© 2002 Nature Publishing Group http://www.nature.

com/naturegenetics contents

supplement september 2002

editorial
Cover art by Darryl Leja
Spreading the word 1
Alan Packer

foreword
Power to the people 2
Andreas D Baxevanis & Francis S Collins

perspective
Genomic empowerment: the importance of public databases 3
Harold Varmus

user’s guide
A user’s guide to the human genome 4
Tyra G Wolfsberg, Kris A Wetterstrand, Mark S Guyer, Francis S Collins
& Andreas D Baxevanis

Introduction: putting it together 5

Question 1 9
How does one find a gene of interest and determine that gene’s structure? Once the
gene has been located on the map, how does one easily examine other genes in that
same region?

Question 2 18
How can sequence-tagged sites within a DNA sequence be identified?

Question 3 21
During a positional cloning project aimed at finding a human disease gene, linkage
data have been obtained suggesting that the gene of interest lies between two
sequence-tagged site markers. How can all the known and predicted candidate genes
in this interval be identified? What BAC clones cover that particular region?

Question 4 29
A user wishes to find all the single nucleotide polymorphisms that lie between two
sequence-tagged sites. Do any of these single nucleotide polymorphisms fall within
the coding region of a gene? Where can any additional information about the
function of these genes be found?

Question 5 33
Given a fragment of mRNA sequence, how would one find where that piece of DNA
mapped in the human genome? Once its position has been determined, how would
one find alternatively spliced transcripts?

40

supplement to nature genetics • september 2002

 contents
Question 6
How would one retrieve the sequence of a gene, along with all annotated exons and
introns, as well as a certain number of flanking bases for use in primer design?

Question 7 44
How would an investigator easily find compiled information describing the structure
of a gene of interest? Is it possible to obtain the sequence of any putative promoter
regions?

Question 8 49
How can one find all the members of a human gene family?

Question 9 53
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Are there ways to customize displays and designate preferences? Can tracks or
features be added to displays by users on the basis of their own research?

Question 10 57
For a given protein, how can one determine whether it contains any functional
domains of interest? What other proteins contain the same functional domains as
this protein? How can one determine whether there is a similarity to other proteins,
not only at the sequence level, but also at the structural level?

Question 11 63
An investigator has identified and cloned a human gene, but no corresponding
mouse ortholog has yet been identified. How can a mouse genomic sequence with
similarity to the human gene sequence be retrieved?

Question 12 66
How does a user find characterized mouse mutants corresponding to human genes?

Question 13 70
A user has identified an interesting phenotype in a mouse model and has been able
to narrow down the critical region for the responsible gene to approximately 0.5 cM.
How does one find the mouse genes in this region?

Commentary: keeping biology in mind 74

Acknowledgments 75

References 76

Web resources: Internet resources featured in this guide 77

supplement to nature genetics • september 2002

© 2002 Nature Publishing Group http://www.nature.com/naturegenetics
supplement
Spreading the word
doi:10.1038/ng961
There was a time, not too long ago, when the wisdom of
genome-sequencing projects was up for discussion.
Would they be too expensive, draining funds from other
areas of the life sciences? Would they be worth the trou-
ble? Not much more than 15 years have passed since
those early debates, and the importance of sequenced
genomes to biology and medicine has now gained wide
acceptance. This is in part owing to the relatively rapid
fall in the cost of sequencing, followed by the undeniably
important insights gained from the annotation of sev-
eral bacterial genomes, and those of a few of our favorite
eukaryotes. The news has been so relentlessly upbeat
that one might even have expected some ‘genome
fatigue’ to set in, especially given the saturation coverage
of the publication of the drafts of the human genome
sequence 18 months ago. Not so, however; witness the
recent jockeying by different groups for inclusion of
‘their’ model organism in the next round of sequencing
projects. The honeymoon goes on.
And yet there are important issues to be addressed.
One is the concern surrounding any bestseller—that it
will have far fewer actual readers than one might expect.
At first glance, this would seem not to apply to the
human genome. After all, one is hard pressed these days
to pick up a copy of Nature Genetics, or any genetics
journal, and not find evidence that sequenced genomes
inform many of the most important advances. A survey
published last year by the Wellcome Trust, however,
found that only half of the researchers who were using
sequence data were fully conversant with the services
provided by the freely accessible databases.
There is also the concern that genome sequencers
might be victims of their own success. As computa-
tional biologist David Roos recently put it, “We are
supplement to nature genetics • september 2002
editorial
september 2002
swimming in a rapidly rising sea of data…how do we
keep from drowning?” And if geneticists and bioinfor-
maticians are struggling to stay afloat, what of the non-
geneticists who are eager to exploit the sequences but
are relative newcomers to the tools needed to navigate
all of this information?
It is with these questions in mind that we present A
User’s Guide to the Human Genome. Written by Tyra
Wolfsberg, Kris Wetterstrand, Mark Guyer, Francis
Collins and Andreas Baxevanis of the National Human
Genome Research Institute (NHGRI), this peer-
reviewed how-to manual guides the reader through
some of the basic tasks facing anyone whose work might
be facilitated by an improved understanding of the
online resources that make sense of annotated genomes.
The directors of these online resources—Ewan Birney of
Ensembl, David Haussler of the University of California,
Santa Cruz and David Lipman of the National Center for
Biotechnology Information—have served as advisors
during the development of this guide, ensuring a bal-
anced and accurate treatment of their respective web
portals. The online version of the guide will also evolve,
with an initial update scheduled for April, 2003.
As noted by Harold Varmus in his eloquent perspec-
tive on A User’s Guide and the public databases it exam-
ines, one of the important legacies of the Human
Genome Project is its ethos of open access to the data. In
this spirit, and with the generous sponsorship of the
NHGRI and the Wellcome Trust, the online version of
this supplement will be freely available on the
Nature Genetics website.
Alan Packer
Nature Genetics
1
 foreword
Power to the people
doi:10.1038/ng962
The National Human Genome Research Institute of the
National Institutes of Health is delighted to sponsor this
special supplement of Nature Genetics. The primary aim
of this supplement is to provide the reader with an ele-
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics
mentary, hands-on guide for browsing and analyzing
data produced by the International Human Genome
Sequencing Consortium, as well as data found in other
publicly available genome databases. The majority of this
supplement is devoted to a series of worked examples,
providing an overview of the types of data available and
highlighting the most common types of questions that
can be asked by searching and analyzing genomic data-
bases. These examples, which have been set in a variety of
biological contexts, provide step-by-step instructions
and strategies for using many of the most commonly-
used tools for sequence-based discovery. It is hoped that
readers will grow in confidence and capability by work-
ing through the examples, understanding the underlying
concepts, and applying the strategies used in the exam-
ples to advance their own research interests.
One of the motivating factors behind the development
of this User’s Guide comes from the general sense that the
most commonly-used tools for genomic analysis still are
terra incognita for the majority of biologists. Despite the
large amount of publicity surrounding the Human
Genome Project, a recent survey conducted on behalf of
2 supplement to nature genetics • september 2002
the Wellcome Trust indicated that only half of biomed-
ical researchers using genome databases are familiar
with the tools that can be used to actually access the data.
The inherent potential underlying all of this sequence-
based data is tremendous, so the importance of all biolo-
gists having the ability to navigate through and cull
important information from these databases cannot be
understated.
The study of biology and medicine has truly undergone
a major transition over the last year, with the public avail-
ability of advanced draft sequences of the genomes of
Homo sapiens and Mus musculus, rapidly growing
sequence data on other organisms, and ready access to a
host of other databases on nucleic acids, proteins and
their properties. Yet for the full benefits of this dramatic
revolution to be felt, all scientists on the planet must be
empowered to use these powerful databases to unravel
longstanding scientific mysteries. As pointed out by
Harold Varmus in the Perspective, free accessibility of all
of this basic information, without restrictions, subscrip-
tion fees or other obstacles, is the most critical component
of realizing this potential. It is our modest hope that this
User’s Guide will provide another useful contribution.
Andreas D. Baxevanis and Francis S. Collins
National Human Genome Research Institute

Genomic empowerment: the importance of
public databases
doi:10.1038/ng963
Over the past twenty five years, a mere sliver of recorded time, the
world of biology — and indeed the world in general — has been
transformed by the technical tools of a field now known as
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics
genomics. These new methods have had at least two kinds of
effects. First, they have allowed scientists to generate extraordi-
narily useful information, including the nucleotide-by-
nucleotide description of the genetic blueprint of many of the
organisms we care about most—many infectious pathogens; use-
ful experimental organisms such as mice, the round worm, the
fruitfly, and two kinds of yeast; and human beings. Second, they
have changed the way science is done: the amount of factual
knowledge has expanded so precipitously that all modern biolo-
gists using genomic methods have become dependent on com-
puter science to store, organize, search, manipulate and retrieve
the new information.
Thus biology has been revolutionized by genomic information
and by the methods that permit useful access to it. Equally
importantly, these revolutionary changes have been dissemi-
nated throughout the scientific community, and spread to other
interested parties, because many of those who practice genomics
have made a concerted effort to ensure that access is simplified
for all, including those who have not been deeply schooled in the
information sciences. The goal of providing genomic informa-
tion widely has also inevitably attracted the interests of those in
the commercial sector, and privately developed versions of vari-
ous genomes are also now available, albeit for a licensing fee.
The operative principle most prominently involved in trans-
mitting the fruits of genomics—the one that has captured the
imagination of the public and served as a standard for the shar-
ing of results and methods more generally in modern biology—
has been open access. Funding by public and philanthropic
organizations, such as the U.S. National Institutes of Health, the
U.S. Department of Energy, the Wellcome Trust in Britain, and
many other organizations, has made this altruistic behavior pos-
sible and has fostered the idea that genomic information about
biological species should be available to all. (Such information
about individual human beings is, of course, an entirely different
matter and should be protected by privacy rules.) The attitude of
open access to new biological knowledge has also been embodied
in the databases of the International Nucleotide Sequence Data-
base Collaboration, comprising the DNA DataBank of Japan, the
European Molecular Biology Laboratory, and GenBank at the US
National Library of Medicine. The same focus on open access is
exemplified by PubMed (operated by the NLM), other gateways
to the scientific literature, and the assemblies of genomic
sequence now found at the several Web portals described in this
guide.
The Human Genome Project (HGP), which has supported the
public genome sequencing effort, has been the mainstay of the
effort to make genomes accessible to the entire community of
scientists and all citizens. This effort has, in fact, been quite natu-
rally extended to instruct the public about many themes in mod-
ern biological science. This has occurred in part because the
human genome itself has been such an exciting concept for the
public; in part because genomes are natural entry points for
supplement to nature genetics • september 2002
perspective
teaching many of the principles of biological design, including
evolution, gene organization and expression, organismal devel-
opment, and disease; and in part because those who work on
genomes have been tireless in attempts to explain the meaning of
genes to an eager public. Endless metaphors, artistic creations,
lively journalism, monographs about social and ethical implica-
tions, televised lectures from the White House, and many other
cultural happenings have been among the manifestations of this
fascination. In this way, the HGP has had a strong hand in raising
the public’s awareness of new ideas in biology and of the power-
ful implications of genomics in medicine, law and other societal
institutions.
Some of these cultural effects come as much from the behav-
ioral aspects of the HGP as from the genomic sequences them-
selves. The sharing of new information, even before its assembly
into publishable form, has spurred efforts to share other kinds of
research tools and has encouraged the notion of making the sci-
entific literature freely accessible through the Internet. The con-
tribution of scientists in many countries to the sequencing of
many genomes, including the human genome, has inspired
efforts to develop gene-based sciences—from basic genomics to
biotechnology—throughout the world, including the poorest
developing nations. Indeed, the World Health Organization, the
United Nations, and the World Bank have all contributed
recently to the growth of the ideas that science is both possible
and valuable in all economies and that science can be a means to
help unify the world’s population under a banner of enlighten-
ment, demonstrating a virtue of globalization.
From this perspective, the availability of the sequences of many
genomes through the Internet is a liberating notion, making
extraordinary amounts of essential information freely accessible
to anyone with a desktop computer and a link to the World Wide
Web. But the information itself is not enough to allow efficient
use. Interested people who reside outside the centers for studying
genomes need to be told where best to view the information in a
form suitable for their purposes and how to take advantage of the
software that has been provided for retrieval and analysis.
The manual before us now offers such help to those who might
otherwise have had trouble in attempting to use the products of
genomics. Furthermore, the advice is offered in that spirit of
altruism that has come to characterize the public world of
genomics. The information is provided in a highly inviting and
understandable format by casting it in the form of answers to the
questions most commonly posed when approaching big
genomes. The information, made freely available on the World
Wide Web, has been assembled by some of the best minds in the
HGP, who have generously given their time and intellect to
encourage widespread use of the great bounty that has been cre-
ated over the past two decades.
In other words, the guide to use of genomes provided here is
simply another indication that the HGP should take great pride
in much more than the sequencing of genomes.
Harold Varmus
Memorial Sloan-Kettering Cancer Center
3
 user’s guide

A user’s guide to the human genome

doi:10.1038/ng964

The primary aim of A User’s Guide to the Human Genome is to provide the reader with an elementary hands-on
guide for browsing and analyzing data produced by the International Human Genome Sequencing Consortium
and other systematic sequencing efforts. The majority of this supplement is devoted to a series of worked exam-
ples, providing an overview of the types of data available, details on how these data can be browsed, and step-
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

by-step instructions for using many of the most commonly-used tools for sequence-based discovery. The major
web portals featured throughout include the National Center for Biotechnology Information Map Viewer, the
University of California, Santa Cruz Genome Browser, and the European Bioinformatics Institute’s Ensembl system,
along with many others that are discussed in the individual examples. It is hoped that readers will become more
familiar with these resources, allowing them to apply the strategies used in the examples to advance their own
research programs.

Authors
Tyra G. Wolfsberg
Kris A. Wetterstrand
Mark S. Guyer
Francis S. Collins
Andreas D. Baxevanis

National Human Genome Research Institute, National Institutes of Health, Bethesda, Maryland 20892, USA.
e-mail: andy@nhgri.nih.gov

4 supplement to nature genetics • september 2002


Introduction: putting it together
doi:10.1038/ng965
In its short history, the Human Genome Project (HGP) has pro-
vided significant advances in the understanding of gene structure
and organization, genetic variation, comparative genomics and
appreciation of the ethical, legal and social issues surrounding
the availability of human sequence data. One of the most signifi-
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics
cant milestones in the history of this project was met in February
2001 with the announcement and publication of the draft ver-
sion of the human genome sequence1. The significance of this
milestone cannot be understated, as it firmly marks the entrance
of modern biology into the genome era (and not the post-
genome era, as many have stated). The potential usefulness of
this rich databank of information should not be lost on any biol-
ogist: it provides the basis for ‘sequence-based biology’, whereby
sequence data can be used more effectively to design and inter-
pret experiments at the bench. The intelligent use of sequence
data from humans and model organisms, along with recent tech-
nological innovation fostered by the HGP, will lead to important
advances in the understanding of diseases and disorders having a
genetic basis and, more importantly, in how health care is deliv-
ered from this point forward2.
Although this flood of data has enormous potential, many
investigators whose research programs stand to benefit in a tan-
gible way from the availability of this information have not
been able to capitalize on its potential. Some have found the
data difficult to use, particularly with respect to incomplete
human genome draft sequence information. Others are simply
not sufficiently conversant with the seeming myriad of data-
bases and analytical tools that have arisen over the last several
years. To assist investigators and students in navigating this
rapidly expanding information space, numerous World Wide
Web sites, courses and textbooks have become available; many
individuals, of course, also turn to their friends and colleagues
for guidance. We have prepared this Guide in that same spirit,
as an additional resource for our fellow scientists who wish to
make use (or better use) of both sequence data and the major
tools that can be used to view these data. The Guide has been
written in a practical, question-and-answer format, with step-
by-step instructions on how to approach a representative set of
problems using publicly available resources. The reader is
encouraged to work through the examples, as this is the best
way to truly learn how to navigate the resources covered and
become comfortable using them on a regular basis. We suggest
that readers keep copies of the Guide next to their computers as
an easy-to-use reference.
Before embarking on this new adventure, it is important to
review a number of basic concepts regarding the generation of
human genome sequence data. This review does not discuss the
chronological development of the HGP or provide an in-depth
treatment of its implications; the reader is referred to Nature’s
Genome Gateway (http://www.nature.com/genomics/human/)
for more information on these topics.
Current status of human genome sequencing
Sequencing of the human genome is nearing completion. The
target date for making the complete, high-accuracy sequence
available is April 2003, the 50th anniversary of the discovery
of the double helix3. As we go to press, however, the work is still
a mosaic of finished and draft sequence. A sequence becomes
supplement to nature genetics • september 2002
user’s guide
finished when it has been determined at an accuracy of at least
99.99% and has no gaps. Sequence data that fall short of that
benchmark but can be positioned along the physical map of the
chromosomes are termed ‘draft’. Currently, 87% of the euchro-
matic fraction of the genome is finished and less than 13% is at
the draft stage.
Even in this incomplete state, the available data are extremely
useful. This usefulness was apparent early on, leading the Inter-
national Human Genome Sequencing Consortium (IHGSC) to
pursue a staged approach in sequencing the human genome. The
first stage generated draft sequence across the entire genome1.
The project is now well advanced into its second stage, with draft
sequence being improved to ‘finished quality’ across the entire
genome, a necessarily localized process. As a result, and as it has
been presented to date, the human genome sequence is an evolv-
ing mix of both finished and unfinished regions, with the unfin-
ished regions varying in data quality. As the data are initially
made available in raw form, with subsequent refinement and
improvement, and because data of different quality are found in
different places in the genome, users must understand the kinds
of data presented by the various tools available.
Determining the human sequence: a brief overview
As with all systematic sequencing projects, the basic experimen-
tal problem in sequencing lies in the fact that the output of a sin-
gle reaction (a ‘read’) yields about 500–800 bp1,4. To determine
the sequence of a DNA molecule that is millions of bases long, it
must first be fragmented into pieces that are within an order of
magnitude of the read size. The sequence at one or both ends of
many such fragments is determined, and the pieces are then
‘assembled’ back into the long linear string from which they were
originally derived. A number of approaches for doing this have
been suggested and tested; the most commonly used is shotgun
sequencing4. The application of shotgun sequencing to the mul-
timegabase- or gigabase-sized genomes of metazoans is still
evolving. A small number of strategies are currently being evalu-
ated, for example, hierarchical or map-based shotgun sequenc-
ing, whole-genome shotgun sequencing and hybrid approaches.
These approaches are described in detail elsewhere4.
The IHGSC’s human sequencing effort began as a purely map-
based strategy and evolved into a hybrid strategy1. The ‘pipeline’
that the IHGSC used to generate the human sequence data
involved the following steps.
1. Bacterial artificial chromosome (BAC) clones were selected,
and a random subclone library was constructed for each one in
either an M13- or a plasmid-based vector.
2. A small number of members of the subclone library (usually
96 or 192) were sequenced to produce very-low-coverage, single-
pass or ‘phase 0’ data. These data were used for quality control
and can be found in the Genome Survey Sequence division of
The DNA Database of Japan (DDBJ), the European Molecular
Biology Laboratory (EMBL) and GenBank (of the National Cen-
ter for Biotechnology and Information; NCBI).
3. If a BAC clone met the requisite standard, subclones were
derived and sufficient sequence data generated from these to pro-
vide four- to fivefold coverage (that is, enough data to represent
an average base in the BAC clone between four and five times).
This is known as ‘draft-level’ coverage, and permits the assembly
5
 user’s guide
5. Subsequent to the genera-
NCBI reference sequences tion and publication of the
The data release and distribution practices adopted by the HGP participants have led not draft human genome sequence,
only to very early, pre-publication access to this treasure trove of information, but also to a work has continued towards
potentially confusing variety of formats and sources for the sequence data. To address this and finishing the sequencing. The
other issues, the NCBI initiated the RefSeq project (http://www.ncbi.nlm.nih.gov/ final stage initially targeted
locuslink/refseq.html). draft-quality BAC clones. For
The goal of the RefSeq effort is to provide a single reference sequence for each molecule of the each of these clones, enough
central dogma: DNA, the mRNA transcript, and the protein. The RefSeq project helps to sim- additional shotgun sequence
plify the redundant information in GenBank by providing, for example, a single reference for data are obtained to bring the
human glyceraldehyde-3-phosphate dehydrogenase mRNA and protein, out of the 14 or so full- coverage to eight- to tenfold, a
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

length sequences in GenBank. Each alternatively spliced transcript is represented by its own ref- stage referred to as ‘fully
erence mRNA and protein. The RefSeq project also includes sequences of complete genomes topped-up’. The data from each
and whole chromosomes, and genomic sequence contigs. The human genomic contigs that fully topped-up BAC are
NCBI assembles, which form the basis of the presentations in the different genome browsers, reassembled, typically resulting
are part of the RefSeq project. Most RefSeq entries are considered provisional and are derived by in a smaller number of contigs
an automated process from existing GenBank records. Reviewed RefSeq entries are manually (often in just a single contig)
curated and list additional publications, gene function summaries and sometimes sequence than at the draft level. The new
corrections or extensions. assembly is again submitted to
Reference sequences are available through NCBI resources, including Entrez, BLAST and the HTGS division as an
LocusLink. They can be easily recognized by the distinctive style of their accession numbers. update of the existing BAC
NM_###### is used to designate mRNAs, NP_###### to designate proteins and NT_###### to clone, now identified with the
designate genomic contigs. The NCBI and UCSC use alignments of the mRNA RefSeqs with the keyword ‘htgs_fulltop’. The
genome to annotate the positions of known genes. Ensembl aligns mRNA RefSeqs to the accession number of the clone
genome. The NCBI also provides model mRNA RefSeqs produced from genome annotation. stays the same, and the version
These are derived by aligning the NM_ mRNAs and other GenBank mRNAs to the assembled number increases by one
genome and then extracting the genomic sequence corresponding to the transcripts. The result- (AC108475.2, for example,
ing model mRNA and model protein sequences have accession numbers of the form becoming AC108475.3).
XM_###### and XP_######. As the XM_ and XP_ records are derived from genomic sequence, 6. At this stage, there are,
they may differ from the original NM_ or GenBank mRNAs because of real-sequence polymor- even for clones comprising a
phisms, errors in the genomic or mRNA sequences or problems in the mRNA/genomic single contig, typically some
sequence alignment. A complete list of types of RefSeqs, along with details on how they are pro- regions that are of insufficient
duced, is available from http://www.ncbi.nlm.nih.gov/LocusLink/refseq.html. quality for the clone to be con-
sidered finished. If this is the
case, the fully topped-up
of sequence using computer programs that can detect overlaps sequence is analyzed by a sequence finisher (an actual person)
between the random reads from the subclones, yielding longer who collects, in a directed manner, the additional data that are
‘sequence contigs’. At this stage, the sequence of a BAC clone needed to close the few remaining gaps and to bring any regions
could typically exist on between four and ten different contigs, of low quality up to the finished sequence standard. While the
only some of which were ordered and oriented with respect to clone is worked on by the finisher, the HTGS entry in GenBank is
one another. The BAC ‘projects’ were submitted, within 24 hours identified by the keyword ‘htgs_activefin’. Once work on the
of having been assembled, to the High-Throughput Genomic clone has been completed, the keyword of the HTG record is
Sequences (HTGS) division of DDBJ/EMBL/GenBank5, where changed to ‘htgs_phase3’, the version number is once again
each was given a unique accession number and identified with increased, and the record is moved from the HTGS division to
the keyword ‘htgs_draft’. (The DDBJ, EMBL and GenBank are the primate division of DDBJ/EMBL/GenBank. In the context of
members of the International Nucleotide Sequence Database a BLAST search at NCBI, these finished BAC sequences would
Collaboration, whose members exchange data nightly and assure now be available in the nr (“non-redundant”) database.
that the sequence data generated by all public sequencing efforts 7. The finished clone sequences are then put together into a
are made available to all interested parties freely and in a timely finished chromosome sequence. As with the initial draft assem-
fashion.) Less-complete high-throughput genomic (HTG) blies, there are a number of steps involved in this process that use
records are also known as ‘phase 1’ records. As the sequence is map-based and sequence-based information in calculating the
refined, it is designated ‘phase 2’. In the context of a BLAST maps. The final assembly process involves identifying overlaps
search at the NCBI, these sequences would be available in the between the clones and then anchoring the finished sequence
HTGS database. contigs to the map of the genome; details of the process can be
4. In late 2000, the draft sequence of the entire human genome found on the NCBI web site (http://www.ncbi.nlm.nih.gov/
was assembled from the sequence of 30,445 clones (BAC clones genome/guide/build.html).
and a relatively small number of other large-insert clones). This Initially, both the UCSC and NCBI groups generated complete
assembled draft human genome sequence was published in Feb- assemblies of the human genome, albeit using different
ruary 2001 and made publicly available through three primary approaches. As noted on the UCSC web site, the NCBI assembly
portals: the University of California, Santa Cruz (UCSC), tended to have slightly better local order and orientation, whereas
Ensembl (of the European Bioinformatics Institute; EBI) and the the UCSC assembly tended to track the chromosome-level maps
NCBI. The use of all three of these sites to obtain annotated somewhat better. Rather than having different assemblies based
information on the human genome sequence is the primary sub- on the same data, IHGSC, UCSC, Ensembl and NCBI decided
ject of this guide. that it would be more productive (and obviously less confusing)

6 supplement to nature genetics • september 2002

 user’s guide
to focus their efforts on a single, definitive assembly. To this end, Over the next year, sequence producers will continue to add
and by agreement, the NCBI assembly will be taken as the refer- finished sequence to the nucleotide sequence databases, and the
ence human genome sequence. It is this NCBI assembly that is NCBI will continue to update the human sequence assembly
displayed at the three major portals covered in this guide. until its ultimate completion. The human genome sequence will,
however, continue to improve even after April 2003, as new
Annotating the assemblies cloning, mapping and sequencing technologies lead to the clo-
Once the assemblies have been constructed, the DNA sequence sure of the few gaps that will remain in the euchromatic regions.
undergoes a process known as annotation, in which useful It is hoped that such technological advances will also allow for
sequence features and other relevant experimental data are cou- the sequencing of heterochromatic regions, regions that cannot
pled to the assembly. The most obvious annotation is that of be cloned or sequenced using currently available methods.
known genes. In the case of NCBI, known genes are identified by The sequence-based and functional annotations presented at
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

simply aligning Reference Sequence (RefSeq) mRNAs (see box), the three major genome portals will certainly continue to evolve
GenBank mRNAs, or both to the assembly. If the RefSeq or Gen- long after April 2003. Computational annotation is a highly
Bank mRNA aligns to more than one location, the best align- active area of research, yielding better methods for identifying
ment is selected. If, however, the alignments are of the same coding regions, noncoding transcribed regions and noncoding,
quality, both are marked on to the contig, subject to certain rules non-transcribed functional elements contained within the
(specifically, the transcript alignment must be at least 95% iden- human sequence.
tical, with the aligned region covering 50% or more of the length,
or at least 1,000 bases). Transcript models are used to refine the Accessing human genome sequence data
alignments. Ensembl identifies ‘best in genome’ positions for Although each of the three portals through which users access
known genes by performing alignments between all known genome data has its own distinctive features, coordination
human proteins in the SPTREMBL database6 and the assembly among the three ensures that the most recent version and anno-
using a fast protein-to-DNA sequence matcher7. UCSC predicts tations of the human genome sequence are available.
the location of known genes and human mRNAs by aligning Ref- Ensembl (http://www.ensembl.org) is the product of a collab-
Seq and other GenBank mRNAs to the genome using the BLAST- orative effort between the Wellcome Trust Sanger Institute and
like alignment tool (BLAT) program8. In addition to identifying EMBL’s European Bioinformatics Institute and provides a bioin-
and placing known genes onto the assemblies, all of the major formatics framework to organize biology around the sequences
genome browser sites provide ab initio gene predictions, using a of large genomes7. It contains comprehensive human genome
variety of prediction programs and approaches. annotation through ab initio gene prediction, as well as infor-
Genome annotation goes well beyond noting where known mation on putative gene function and expression. The web site
and predicted genes are. Features found in the Ensembl, NCBI provides numerous different views of the data, which can be
and UCSC assemblies include, for example, the location and either map-, gene- or protein-centric. Ensembl is actively build-
placement of single-nucleotide polymorphisms, sequence- ing comparative genome sequence views, and presents data
tagged sites, expressed sequence tags, repetitive elements and from human, mouse, mosquito and zebrafish. In addition,
clones. Full details on the types of annotation available and the numerous sequence-based search tools are available, and the
methods underlying sequence annotation for each of these dif- Ensembl system itself can be downloaded for use with individ-
ferent types of sequence feature can be found by accessing the ual sequencing projects.
URLs listed under Genome Annotation in the Web Resources The UCSC Genome Browser (http://genome.ucsc.edu) was
section of this guide. At UCSC, many of the annotations are pro- originally developed by a relatively small academic research
vided by outside groups, and there may be a significant delay group that was responsible for the first human genome assem-
between the release of the genome assembly and the annotation blies. The genome can be viewed at any scale and is based on
of certain features. Furthermore, some tracks are generated for the intuitive idea of overlaying ‘tracks’ onto the human
only a limited number of assemblies. For an in-depth discussion genome sequence; these annotation tracks include, for exam-
of genome annotation, the reader is referred to an excellent ple, known genes, predicted genes and possible patterns of
review by Stein9 and the references cited therein. This review, alternative splicing. There is also an emphasis on comparative
along with the Commentary in this guide, also provides cautions genomics, with mouse genomic alignments being available.
on the possible overinterpretation of genome annotation data. The browser also provides access to an interactive version of
the BLAT algorithm8, which UCSC uses for RNA and compar-
The data—and sometimes the tools—change every day ative genomic alignments.
The steps outlined in the previous section should emphasize Given its Congressional mandate to store and analyze biologi-
that the state of the human genome sequence will continue to be cal data and to facilitate the use of databases by the research com-
in flux, as it will be updated daily until it has actually been munity, the NCBI (http://www.ncbi.nlm.nih.gov) serves as a
declared ‘finished’. (Finished sequence is properly defined as the central hub for genome-related resources. NCBI maintains Gen-
“complete sequence of a clone or genome, with an accuracy of at Bank, which stores sequence data, including that generated by
least 99.99% and no gaps”2. A more practical definition is that of the HGP and other systematic sequencing projects. NCBI’s Map
“essentially finished sequence,” meaning the complete sequence Viewer provides a tool through which information such as exper-
of a clone or genome, with an accuracy of at least 99.99% and no imentally verified genes, predicted genes, genomic markers,
gaps, except those that cannot be closed by any current physical maps, genetic maps and sequence variation data can be
method.) The reader should be mindful of this, not just when visualized. The Map Viewer is linked to other NCBI tools—for
reading this guide, but also, when referring back to it over time. example, Entrez, the integrated information retrieval system that
Similarly, the tools used to search, visualize and analyze these provides access to numerous component databases.
sequence data also undergo constant evolution, capitalizing on Although we have chosen to illustrate each example using
new knowledge and new technology in increasing the usefulness resources available at a single site, almost all the questions in this
of these data to the user. guide can be answered using any of the three browsers. The

supplement to nature genetics • september 2002 7

 user’s guide
informational sidebars that follow some of the questions provide
pointers on how to format the search at other sites. Furthermore,
the three sites link to each other wherever possible. Examples
presented in this Guide rely on the data and genome browser
interfaces that were available in June 2002. As new versions of the
genome assembly and viewing tools will come online every few
months, the specifics of some of the examples may change over
time. Regardless, the basic strategies behind answering the ques-
tions in the examples will remain the same. This underscores the
importance of readers working through the examples at their
own computers so that they may understand and be able to navi-
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics
gate these public databases. The readers are encouraged to
explore the alternative methods for answering the questions.
8 supplement to nature genetics • september 2002
Browser problems?
In following the question-and-answer portion of this guide,
some readers may find that their web browsers are not be able
to render the web pages properly. If this occurs, do one or
more of the following:
1. Install the most recent version of either Netscape Navi-
gator or Internet Explorer.
2. Increase the amount of memory available to the web
browser.
3. Try a different web browser. In general, Macintosh users
who seek to gain access to these three genome portals will see
better performance with Internet Explorer.

Question 1
How does one find a gene of interest and determine that gene’s struc-
ture? Once the gene has been located on the map, how does one easily
examine other genes in that same region?
doi:10.1038/ng966
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics
This question serves as a basic introduction to the three major
genome viewers. One gene, ADAM2, will be examined using
all three sites so that the reader can gain an appreciation of
the subtle differences in information presented at each of
these sites.
National Center for Biotechnology Information Map
Viewer
The NCBI Human Map Viewer can be accessed from the NCBI’s
home page, at http://www.ncbi.nlm.nih.gov. Follow the hyper-
link in the right-hand column labeled Human map viewer to go
to the Map Viewer home page. The notation at the top of the
page indicates that this is Build 29, or the NCBI’s 29th assembly
of the human genome. Build 29 is based on sequence data from 5
April 2002. The previous genome assembly, Build 28, was based
on sequence data from 24 December 2001. To search for any
mapped element, such as a gene symbol, GenBank accession
number, marker name or disease name, enter that term in the
Search for box and then press Find. For this example, enter
‘ADAM2’ and then press Find. The on chromosome(s) box may be
left blank for text-based searches such as this one.
The resulting overview page shows a schematic of all of the
human chromosomes, pinpointing the position of ADAM2 to
the p arm of chromosome 8 (Fig. 1.1). The search results section
shows that the gene exists on two NCBI maps, Genes_cyto and
Genes_seq. Genes_cyto refers to the cytogenetic map, whereas
Genes_seq refers to the sequence map. Clicking on either of those
two links opens a view of just that map.
Detailed descriptions of these and other NCBI maps are
available at http://www.ncbi.nlm.nih.gov/PMGifs/Genomes/
humansearch.html. To get the most general overview of the
genomic context of ADAM2, including all available maps, click
on the item in the Map element column (in this case, ADAM2).
This view shows ADAM2 and a bit of flanking sequence on chro-
mosome 8p11.2 (Fig. 1.2). Three maps are displayed in this view,
each of which will be discussed below. Additional maps, dis-
cussed in other examples in this guide, can be added to this view
using the Maps & Options link.
The rightmost map is the master map, the map providing the
most detail. The master map in this case is the Genes_seq map,
which depicts the intron/exon organization of ADAM2 and is
created by aligning the ADAM2 mRNA to the genome. The gene
appears to have 14 exons. The vertical arrow next to the ADAM2
gene symbol (within the pink box) shows the direction in which
the gene is transcribed. The gene symbol itself is linked to
LocusLink, an NCBI resource that provides comprehensive
information about the gene, including aliases, nucleotide and
protein sequences, and links to other resources10 (see Question
10). The links to the right of the gene symbol point to additional
information about the gene.
• sv, or sequence view, shows the position of the gene in the
context of the genomic contig, including the nucleotide and
encoded protein sequences.
supplement to nature genetics • september 2002
user’s guide
• ev brings the user to the evidence viewer, a view that displays
the biological evidence supporting a particular gene model.
This view shows all RefSeq models, GenBank mRNAs, tran-
scripts (whether annotated, known or potential) and
expressed sequence tags (ESTs) aligning to this genomic con-
tig. More information on the evidence viewer can be found
on the NCBI web site by clicking Evidence Viewer Help on any
ev report page.
• hm is a link to the NCBI’s Human–Mouse Homology Map,
showing genome sequences with predicted orthology
between mouse and human (Fig. 12.2).
• seq allows the user to retrieve the genomic sequence of the
region in text format. The region of sequence displayed can
easily be changed.
• mm is a link to the Model Maker, which shows the exons that
result when GenBank mRNAs, ESTs and gene predictions are
aligned to the genomic sequence. The user can then select
individual exons to create a customized model of the gene.
More information on the Model Maker can be found on the
NCBI web site by clicking help on any mm report page.
The UniG_Hs map shows human UniGene clusters that have
been aligned to the genome. The gray histogram depicts the
number of aligning ESTs and the blue lines show the mapping of
UniGene clusters to the genome. The thick blue bars are regions
of alignment (that is, exons) and the thin blue lines indicate
potential introns. In this example, the mapping of UniGene clus-
ter Hs.177959 to the genome follows that of ADAM2, and all the
exons align.
The Genes_cyto map shows genes that have been mapped
cytogenetically; the orange bar shows the position of the gene.
Although ADAM2 has been finely mapped and is represented by
a short line, other genes, such as the group below it on a longer
line, have been cytogenetically mapped to broader regions of
chromosome 8.
Clicking on the zoom control in the blue sidebar allows the
user to zoom out to view a larger region of chromosome 8.
Zooming out one level shows 1/100th of the chromosome. There
are 20 genes in the region, and all 20 are labeled (displayed) in
this view (Fig. 1.3). The region of ADAM2 is highlighted in red
on all maps. On the basis of the Genes_seq map, ADAM2 is
located between ADAM18 and LOC206849.
University of California, Santa Cruz Genome Browser
The home page for the UCSC Genome Browser is http://genome.
ucsc.edu/. At present, UCSC provides browsers not only for the
most recent version of the mouse and human genome data, but
also for several earlier assemblies. To use the Genome Browser,
select the appropriate organism from the pull-down menu at the
top of the blue sidebar (Human, in this case) and then click the
link labeled Browser. On the resulting page, select the version of
the human assembly to view. The genome browser from August
2001 is based on an assembly of the human genome done by
UCSC using sequence data available on that date. The Dec. 2001
9
 user’s guide
browser displays annotations based on NCBI’s build 28 of the
human genome, and the Apr. 2002 browser displays annotations
on NCBI’s build 29. As the annotations presented in this most
recent human assembly are not yet as comprehensive as those
from the December 2001 assembly, the examples in this text are
based on the earlier assembly. Select Dec. 2001 from the pull-
down menu to access the assembly from that date (Fig. 1.4).
Supported types of queries are listed below the text input
boxes. Enter ‘ADAM2’ in the box labeled position and then
click Submit. The results of this search are presented in two
categories, Known Genes and mRNA Associated Search Results
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics
(Fig. 1.5). The section marked Known Genes shows the map-
ping of the NCBI Reference mRNA sequences to the genome.
The mRNA Associated Search Results represent the mapping of
other GenBank mRNA sequences to the genome. Click on the
Known Genes link for ADAM2 (arrow, Fig. 1.5) to see the
genomic context of the ADAM2 mRNA Reference Sequence
(NM_001464).
The resulting zoomed-in view shows a region of chromosome
8 from base pair 36234934 to 36280132, located within 8p12
(Fig. 1.6). The blue track entitled Known Genes (from RefSeq)
shows the intron–exon structure of known genes. The vertical
boxes indicate exons and the horizontal lines introns. The
ADAM2 gene seems to have 14 exons. The direction of transcrip-
tion is indicated by the arrowheads on the introns. The tracks
labeled Acembly Gene Predictions, Ensembl Gene Predictions
and Fgenesh++ Gene Predictions are the results of gene predic-
tions (see Question 7). Alignments of other database nucleotide
sequences are shown in the Human mRNAs from GenBank,
spliced EST, UniGene and Nonhuman mRNAs from GenBank
tracks. Translated alignments of mouse and Tetraodon genomic
sequence are in the mouse and fish BLAT tracks. Tracks display-
ing single-nucleotide polymorphisms (SNPs), repetitive ele-
ments and microarray data are shown at the bottom. Additional
details about each track are available by selecting the track name
in the Track Controls at the bottom.
To view the genomic context of ADAM2, zoom out 10× by
clicking on the zoom out 10× box in the upper right corner.
ADAM2 is located between TEM5 and ADAM18 (Fig. 1.7).
Ensembl
The Ensembl7 project, http://www.ensembl.org/, provides
genome browsers for four species: human, mouse, zebrafish and
mosquito. Click on Human to view the main entry point for the
human genome. The current version of human Ensembl is ver-
sion 6.28.1, based on the NCBI’s 28th build of the genome. To
perform a text search, enter ‘ADAM2’ in the text box, and limit
the search by selecting Gene from the pull-down search. Click on
the upper button labeled Lookup. A single result is returned with
a link to the ADAM2 gene (Fig. 1.8).
Click on either of the ADAM2 links to retrieve the GeneView
window. The returned page contains four sections of data. The
first section (Fig. 1.9) is an overview of ADAM2, including links
to accession numbers and protein domains and families. Links to
the Ensembl view of highly similar mouse sequences are pre-
sented in the Homology Matches section. Some of these fields will
be described in more detail in later examples. The second section
of the GeneView window provides information on the gene tran-
script (Fig. 1.10). The sequence of the cDNA is shown, as is a
graphic of its intron–exon structure. A limited amount of the
genomic context around the gene is shown schematically as well.
10
Exon sequences are shown in the third section of the GeneView
(Fig. 1.11) and splice sites in the fourth (Fig. 1.12). If more than
one transcript is predicted for the gene, each is allocated its own
transcript, exon and splice-site sections.
The complete genomic context of ADAM2 is viewed by return-
ing to the first section of the GeneView (Fig. 1.9) and clicking on
one of the two links within the Genomic Location box. The top
portion of the resulting ContigView (Fig. 1.13) depicts the chro-
mosome, with the region of interest outlined in red. The
Overview shows the genomic context of the gene, including the
chromosome bands, contigs, markers and genes that map to near
8p12. Clicking on any of these items recenters the display around
that item. The section of interest is boxed in red on the
DNA(contigs) map. The genes annotated by Ensembl as being
around ADAM2 are Q96KB2 and ADAM18.
The bottom panel of the ContigView, the Detailed View
(Fig. 1.14), shows a zoomed-in view of the boxed region, high-
lighting all features that have been mapped to this region of the
human genome. The navigator buttons between the Overview
and the Detailed View move the display to the left and right and
zoom in and out. The features to be displayed can be changed
by selecting the Features pull-down menu and then checking
which features to view.
The Features shown in Fig. 1.14 are the defaults. The DNA
(contigs) map separates items on the forward strand (above)
from those on the reverse (below). The only feature on the
reverse strand in this view is a single Genscan transcript, pre-
dicted by the GENSCAN gene prediction program11 (see Ques-
tion 7). The forward strand shows five types of features. Starting
at the bottom, the ADAM2 transcript is shown in red, indicating
that it is a known transcript corresponding to a near-full-length
cDNA sequence, protein sequence or both already available in
the public sequence database. Black transcripts are predicted
based on EST or protein sequence similarity. EST Transcr. links to
individual aligning ESTs, whereas the UniGene track near the top
displays UniGene clusters. The Genscan model on the forward
strand contains many exons found in the known transcript. The
Proteins and Human proteins boxes indicate protein sequences
that align to this version of the genome, whereas NCBI Transcr.
links to the NCBI Map Viewer. Positioning the computer mouse
over any feature brings up the feature’s name and links to more
detailed information.
The NCBI, UCSC and Ensembl sometimes use different sym-
bols for the same genes, so it can be difficult to compare the
views obtained by the different browsers. Furthermore, the
three sites maintain independent annotation pipelines and do
not all attempt to align the same mRNA sequences to the
genome. The NCBI is currently displaying build 29, Ensembl
shows build 28, and UCSC offers both builds 28 (December
2001) and 29 (April 2002), although all examples from UCSC in
this guide will be illustrated using the better-annotated build
28. Because of the differences between the two assemblies, there
are subtle discrepancies between what is shown at the NCBI and
what is available at UCSC and Ensembl. However, it is fairly
easy to navigate among the three sites. The NCBI, for example,
links to Ensembl and UCSC through the black boxes at the top
of LocusLink entries for human genes, and Ensembl directs
users to NCBI and UCSC through the “Jump to” link in its Con-
tigView. Some versions of UCSC’s Genome Browser have links
to Ensembl and NCBI’s Map Viewer in the blue bar at the top of
each browser page.
supplement to nature genetics • september 2002
 user’s guide
Figure 1.1
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 1.2

supplement to nature genetics • september 2002 11

 user’s guide
Figure 1.3
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 1.4

12 supplement to nature genetics • september 2002

 user’s guide
Figure 1.5
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 1.6

supplement to nature genetics • september 2002 13

 user’s guide
Figure 1.7
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 1.8

14 supplement to nature genetics • september 2002

 user’s guide
Figure 1.9
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 1.10

supplement to nature genetics • september 2002 15

 user’s guide
Figure 1.11
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 1.12

16 supplement to nature genetics • september 2002

 user’s guide
Figure 1.13
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 1.14

supplement to nature genetics • september 2002 17

 user’s guide
Question 2
How can sequence-tagged sites within a DNA sequence be identified?
doi:10.1038/ng967
The NCBI’s electronic PCR (e-PCR) tool12, which is part of the
UniSTS resource, can be used to find STS markers within a DNA
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics
fragment of interest. UniSTS (http://www.ncbi.nih.gov/
genome/sts/) contains all the available data on STS markers,
including primer sequences, product size, mapping information
and alternative names. Links to other NCBI resources such as
Entrez, LocusLink and the MapViewer are also provided. e-PCR
looks for potential STSs in a DNA sequence by searching for sub-
sequences with the correct orientation and distance that could
represent the PCR primers used to generate known STSs.
The e-PCR home page can be found by going to the NCBI
home page, at http://www.ncbi.nlm.nih.gov, and then following
the Electronic PCR link in the right-hand column. On the e-PCR
home page, paste the sequence of interest or enter an accession
number into the large text box at the top of the page. The acces-
sion number of the sequence for this example is AF288398. This
sequence contains only one STS, stSG47693, which is located
between nucleotides (nt) 2102 and 2232 of the sequence under
study (Fig. 2.1).
Click on the marker name to bring up details of the STS from
UniSTS (Fig. 2.2). The primer information and PCR product size
are listed at the top of the page, along with alternative names for
the marker. Often STSs are known by different names on differ-
18
ent maps. Cross-references to LocusLink, UniGene and the
Genebridge 4 map to which this STS was mapped are shown
next. The mapping information section contains links to the
NCBI’s MapViewer. At the bottom of the page, the Electronic
PCR results show other sequences, including contigs, mRNAs
and ESTs that may contain this STS marker.
To see the genomic context of the STS marker in all maps to
which it has been mapped, click on the link labeled MapViewer
at the top of the Mapping Information section. This map view
(Fig. 2.3) shows two maps. Note that, in this view, the STS
stSG47693 is called RH92759 (highlighted in pink). Gene
Map ’99–Genebridge 4 (GM99_GB4, left) has 46,000 STS mark-
ers mapped onto the GB4 RH panel by the International
Radiation Hybrid Consortium. The STS map (right) shows the
NCBI’s placement of STSs onto the genome sequence assembly
using e-PCR. Gray lines connect markers that appear in both
maps, whereas the red line denotes where the STS RH92759
appears on both maps. In the region shown, there are a total of
211 STSs on the STS map, but only 20 are labeled in this view. To
the right of the STS map, the green and yellow circles show the
maps on which the STS markers have been placed. One can
zoom in or out of this view by clicking on the lines of the zoom
tool in the left sidebar.
supplement to nature genetics • september 2002
 user’s guide
Figure 2.1
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 2.2

supplement to nature genetics • september 2002 19

 user’s guide
Figure 2.3
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

20 supplement to nature genetics • september 2002


Question 3
During a positional cloning project aimed at finding a human disease
gene, linkage data have been obtained suggesting that the gene of
interest lies between two sequence-tagged site markers. How can all
the known and predicted candidate genes in this interval be identified?
What BAC clones cover that particular region?
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics
doi:10.1038/ng968
UCSC
One possible starting point for this search is the UCSC Genome
Browser home page, at http://genome.ucsc.edu. From this page,
select Human from the Organism pull-down menu in the blue
bar at the side of the page, and then click Browser. On the Human
Genome Browser Gateway page, change the assembly pull-down
to Dec. 2001. To view a region of the genome between two query
terms, enter the terms in the search box, separated by a semi-
colon. For example, to view the region between STS markers
D10S1676 and D10S1675, enter ‘D10S1676;D10S1675’ in the box
marked position and press Submit. Because both of these markers
map to a single position in the genome, the genome browser for
the region between those markers is returned (Fig. 3.1).
The STS Markers track displays genetically mapped markers in
blue and radiation hybrid–mapped markers in black. Click on
the STS Markers label to expand that track and see each marker
listed individually (Fig. 3.2). The markers of interest are called by
their alternate names (AFMA232YH9 and AFMA230VA9 in this
view) and are at the top and bottom of the interval, respectively
(Fig. 3.2, arrows).
The full list of known genes in this display is shown in the
Known Genes track (Fig. 3.1). These protein-coding genes are
taken from the RefSeq mRNA sequences compiled at the NCBI10
and aligned to the genome assembly using the BLAT program8. To
export a list of the genes, or other features, in this region, click the
Tables link in the top blue bar. For more information about a par-
ticular gene (such as MGMT), click on the gene symbol to get a list
of additional links to resources such as Online Mendelian Inheri-
tance in Man (OMIM), PubMed, GeneCards and Mouse Genome
Informatics (MGI; Fig. 3.3). Many tracks, including Acembly
Genes, Ensembl Genes and Fgenesh++ Genes, indicate predicted
genes (see Question 7).To view the full set of features in any of
these categories, click on the title of that track on the left side of the
screen in Fig. 3.1. To view brief descriptions of these tracks, as well
as others not mentioned, click on the gray box to the left of the
track or scroll down to Track Controls and click on the title of a fea-
ture of interest. Explanations of the gene-prediction programs can
be found in Question 7. Reset the browser to its default settings by
clicking on the reset all button below the tracks.
To see the BAC clones used for sequencing, return to the page
illustrated in Fig. 3.1 and click on Coverage at the left side of the
screen to expand that track. Here BAC clones are listed individu-
ally, with finished regions shown in black and draft regions
shown in various shades of gray (Fig. 3.4). For details such as size
and sequence coverage of a specific clone, click on the clone
accession number (such as AL355529.21, arrow). From this
screen, click on the accession number (as shown in Fig. 3.5) to
link to the NCBI Entrez document summary for the clone. The
full GenBank entry can be viewed by clicking on AL355529 on
the Entrez document summary page.
supplement to nature genetics • september 2002
user’s guide
According to NCBI naming conventions, this clone is from the
RP11 library and has been named 85C15. RP11 is the NCBI desig-
nation for RPCI-11, a commonly used human BAC library pro-
duced at the Roswell Park Cancer Institute. More information
on the naming conventions of genomic sequencing libraries
can be found at the NCBI’s Clone Registry (Fig. 3.6;
http://www.ncbi.nlm.nih.gov/genome/clone/nomenclature.shtml).
Clone ordering information is also available, at http://www.
ncbi.nlm.nih.gov/genome/clone/ordering.html.
NCBI
The NCBI MapViewer allows for direct viewing of the region
between two markers, as long as both markers are on the master
map. If, for example, the master map is a cytogenetic one, one
can search chromosome 22 for the region between band num-
bers 22q12.1 and 22q13.2. If the master map is Gene_Seq, one
can view the region between two mapped genes.
Access the Map Viewer home page by starting at the NCBI
home page (http://www.ncbi.nlm.nih.gov) and clicking
Human map viewer in the list on the right-hand side of the
page. To view multiple hits on the same chromosome, type in
the search terms separated by the word ‘OR’. To see the same
region between the STS markers D10S1676 and D10S1675, for
example, type ‘D10S1676 OR D10S1675’ in the search box, and
hit Find. At the top of the resulting page (Fig. 3.7), two red tick
marks on the chromosome cartoon indicate that the markers
map close to each other on chromosome 10. The search results
at the bottom of the page show the alternative names for the
two markers (AFMA232YH9 and AFMA230VA9) as well as the
maps on which they have been placed. To view both markers at
the same time, click on the link for chromosome 10 in the
chromosome diagram. Fig. 3.8 shows the region around
D10S1676 and D10S1675, with the original queries high-
lighted in pink. Red lines connect the positions of the marker
on the different maps.
The Maps & Options link, in the horizontal blue bar near the
top of the page, allows the user to customize the maps and region
displayed. To view, for example, the known and predicted genes
One can also search for a region between two STS markers
using the MapView at Ensembl. Start at the Ensembl Human
Genome Browser at http://www.ensembl.org/Homo_sapi-
ens/, click on the idiogram of any chromosome to access the
MapView, and enter the marker names in the Jump to Con-
tigview section. To use Ensembl to obtain a list of genes (or
other annotations) in a defined chromosomal region, click on
Export→Gene List from any ContigView window (Fig. 1.14,
center yellow bar).
21
 user’s guide
in this region, as well as the BAC clones from which the sequence
was derived, click on the link to open the Maps & Options win-
dow (Fig. 3.9). First remove all the maps except Gene and STS
from the Maps Displayed box by highlighting them, and selecting
<<REMOVE. Next, add the Transcript (RNA), GenomeScan,
Component and Contig maps by selecting them from the Avail-
able Maps box and selecting ADD>>. Make the STS map the
master by highlighting it, then selecting Make Master/Move to
Bottom. To limit the view such that only the STSs between
D10S1676 and D10S1675 are shown, type the marker names in
the Region Shown boxes. Hit Apply to see the aligned maps. In
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics
some cases, it may be useful to select a page size larger than the
default of 20 to view more data in the browser window.
Fig. 3.10 shows the maps, as specified in the Maps & Options
window. The green dots to the right of the STS map show all the
maps on which the markers appear. This is a fairly long region of
chromosome 10, and not every STS marker is shown. In particu-
lar, although there are 611 STSs in this region, only 20 are shown
by name in this view. For each known gene, the Genes_Seq map
shows all the exons that have been mapped to the genome. Exons
for individual known mRNAs are shown on the RNA (Tran-
script) map. Unless a gene is alternatively spliced, the Genes_Seq
and RNA maps will be the same. The GScan (GenomeScan) map
22
shows the NCBI’s gene predictions. Any of these genes, known or
predicted, are candidates for the disease gene.
The NCBI’s assembled contigs, also known as the NT contigs,
are found in the Contig map. Blue segments come from finished
sequence, orange from draft. These contigs are constructed from
the individual GenBank sequence entries shown in the Comp
(Component) map. Draft HTG records (phase 1 and 2; see
http://www.ncbi.nlm.nih.gov/HTGS/) are displayed in orange
and finished HTGs in blue. Most of these GenBank entries are
derived from BAC clones. The tiling paths of the BAC clones that
were assembled into contigs are clearly visible. One can obtain
more details about an entry, including the clone name, by click-
ing on the accession number to link to Entrez. The clone name is
visible directly in the MapViewer if the Comp map is the master.
A map can be quickly made the master map by clicking on the
blue arrow next to its name.
Because this is a zoomed-out view of the chromosome, indi-
vidual genes and GenBank entries are difficult to visualize.
Zooming in, using the controls in the blue sidebar, will provide
a region in more detail. Alternatively, click on the Data As
Table View in the left sidebar to retrieve all data, including
those hidden in this view, as a text-based table (partially shown
in Fig. 3.11).
supplement to nature genetics • september 2002
 user’s guide
Figure 3.1
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 3.2

supplement to nature genetics • september 2002 23

 user’s guide
Figure 3.3
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 3.4

24 supplement to nature genetics • september 2002

 user’s guide
Figure 3.5
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 3.6

supplement to nature genetics • september 2002 25

 user’s guide
Figure 3.7
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 3.8

26 supplement to nature genetics • september 2002

 user’s guide
Figure 3.9
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 3.10

supplement to nature genetics • september 2002 27

 user’s guide
Figure 3.11
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

28 supplement to nature genetics • september 2002


Question 4
A user wishes to find all the single nucleotide polymorphisms that lie
between two sequence-tagged sites. Do any of these single nucleotide
polymorphisms fall within the coding region of a gene? Where can any
additional information about the function of these genes be found?
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics
doi:10.1038/ng969
The starting point for this search would be the web site for the
Database of Single Nucleotide Polymorphisms (dbSNP) at the
NCBI13, which is located at http://www.ncbi.nlm.nih.gov/SNP.
There is a series of links on the page that allow the user to search
using either information about the database submission itself or
information regarding genes and gene loci.
For this particular search, assume that the region of interest is
known and defined by two STS markers, RH70674 and G32133.
Begin by scrolling to the section labeled Between Markers at the
bottom of the page. Enter the STS marker names ‘RH70674’ and
‘G32133’ into the two text boxes, and click on Submit STS Mark-
ers. This will produce a display showing SNPs 1–25 out of the
total of 81 within the region of interest. Go to page 3 of the dis-
play by entering ‘3’ in the Page box and clicking Display.
The resulting page (Fig. 4.1) illustrates most of the possible
types of result one would find on a typical dbSNP results page. In
the table, starting from the left, the first column gives the individ-
ual dbSNP cluster IDs (all starting with ‘rs’). The second column,
labeled Map, shows whether a particular SNP has been mapped
to a unique position in the genome (illustrated by a single green
arrow, as in the first row of the example) or to multiple positions
(not shown here).
The next set of columns, labeled Gene, indicates whether these
SNPs are associated with particular features, such as genes,
mRNAs or coding regions. The three columns (L, T and C) are
either lit up or appear gray in every row. Taking each in order:
If the L (for locus) appears in blue, part or all of the marker
position lies either within 2 kilobases (kb) of the 5′ end of a gene
feature or within 500 bases of the 3′ end of a gene feature.
If the T (for transcript) appears in green, part or all of the
marker position overlaps with a known mRNA. This does not
mean, however, that the SNP marker necessarily falls within a
coding region.
If the C (for coding) appears in orange, part or all of the
marker position overlaps with a coding region.
The next column, labeled Het, indicates the average heterozy-
gosity observed for this marker, on a scale of 0–100%. A reading
of zero means that no information is available for that particular
marker, whereas the pink bars show a 95% confidence interval
for the marker. The Validation column indicates whether the
marker has been validated (shown by a star) or is unvalidated
(shown by light blue boxes). Validated markers have been veri-
fied by independent re-analysis of the sequence. All of the unval-
idated markers shown in Fig. 4.1 are denoted by three blue boxes,
which, according to the scale at the top of the column, means that
there is a >95% success rate in validation. This figure indicates
the probability that this marker is real. (The success rate is
defined as 1 – false-positive rate.)
In the penultimate column, the symbol TT (not shown here)
indicates that individual genotypes are available for this marker.
Finally, the Linkout Avail column indicates which markers are
supplement to nature genetics • september 2002
user’s guide
linked to other databases; a P in this column indicates that the
variation has been mapped to a known protein structure. For a
complete description of all the features within this display, click
on any part of the header above the columns.
Returning to the original question, one of the SNPs displayed
on this page does indeed fall within a coding region, as indi-
cated by an orange C. To obtain more information on any par-
ticular SNP, simply click on the hyperlinked SNP Cluster ID.
Clicking on rs1059133, for example, produces a new page, with
all available information on that SNP (Fig. 4.2). Under the
header marked Submitter records for this RefSNP Cluster is a list
of the individual SNPs (in this case, only one SNP) that have
been clustered together to form this single reference SNP. The
sequence of the SNP is shown in the next header. Under the
header marked NCBI Resource Links are GenBank and NCBI
RefSeq entries that are associated with this SNP. Scrolling fur-
ther down on the SNP page (Fig. 4.3), the gene whose coding
region this SNP falls within is indicated on the LocusLink Analy-
sis section (ADAM2, a disintegrin and metalloproteinase
domain 2). The SNP allele is G/C, a non-synonymous change
leading to replacement of the Asp residue in the reference
sequence by a His residue. Links are also provided to the NCBI
Map Viewer, Ensembl map and UCSC genome assembly in the
section labeled Integrated Maps. The sections labeled Variation
Summary and Validation Summary (not shown) give the raw
data on this particular SNP.
To answer the final part of this question requires jumping from
dbSNP to LocusLink10. To do so, click on the ADAM2 link in the
line marked LocusLink at the top of the page (Fig. 4.3). This
brings the user to the LocusLink page for ADAM2 and provides
numerous jumping-off points to the NCBI and affiliated
resources through the boxed links at the top of the page. More
information on these resources can be found by following the
LocusLink FAQ link in the left-hand column of the page. By sim-
ply examining the LocusLink page itself, one sees that the
ADAM2 protein belongs to a family of membrane-anchored pro-
teins that have been implicated in processes as diverse as fertiliza-
tion, muscle development and neurogenesis.
One often-overlooked source of information on genes and
gene products is OMIM14. This is an electronic version of the
Using the UCSC browser, users can retrieve the positions of
genome annotations such as SNPs as a text file suitable for
loading into a spreadsheet program. While looking at the
browser for a defined chromosomal region, click on the
Tables link (Fig. 1.6, upper blue bar). Similarly, to export a
list of genome annotations in a defined chromosomal region
at Ensembl, click on Export from any ContigView window
(Fig. 1.14, center yellow bar).
29
 user’s guide
catalog of human genes and genetic disorders developed by Vic-
tor McKusick at The Johns Hopkins University. OMIM provides
the user with concise textual information from the published
literature on most human disorders with a genetic basis, and
links back to the primary literature as appropriate. Information
comprising an OMIM entry includes the gene symbol, alternate
names for the disease, a description of the disease (including
clinical, biochemical and cytogenetic features), details of the
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics
30
mode of inheritance (including mapping information) and a
clinical synopsis. These entries are manually curated, ensuring
that the ‘executive summary’ is up to date and accurate.
Although OMIM can be searched directly, many LocusLink
entries also link to the OMIM record for the gene. The OMIM
entry page for the ADAM2 protein is shown in Fig. 4.4. The
page is fully hyperlinked to PubMed, GenBank and other
related databases.
supplement to nature genetics • september 2002
 user’s guide
Figure 4.1
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 4.2

supplement to nature genetics • september 2002 31

 user’s guide
Figure 4.3
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 4.4

32 supplement to nature genetics • september 2002


Question 5
Given a fragment of mRNA sequence, how would one find where that
piece of DNA mapped in the human genome? Once its position has been
determined, how would one find alternatively spliced transcripts?
doi:10.1038/ng970
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics
For the purpose of this example, the fragment of mRNA of inter-
est is contained within GenBank accession number BG334944.
First, retrieve the nucleotide sequence of this EST using the
NCBI’s Entrez interface, at http://www.ncbi.nlm.nih.gov/
Entrez/. Type ‘BG334944’ into the text box at the top of the page,
change the pull-down menu to Nucleotide and press Go. The
resulting page shows one entry, corresponding to accession num-
ber BG334944. To retrieve this sequence in FASTA format (a
common format for bioinformatics programs), change the pull-
down menu on this page to FASTA and then press Text (Fig. 5.1).
A new web page containing only the sequence, in FASTA format,
is produced (Fig. 5.2); copy the resulting sequence.
To determine where this sequence maps within the genome,
use UCSC’s BLAT tool8. Begin this search by pointing your web
browser to the UCSC Genome Browser home page, at
http://genome.ucsc.edu. From this page, select Human from the
Organism pull-down menu in the blue bar on the side of the
page, and then click Blat. Paste the FASTA-formatted sequence
obtained from Entrez (above) into the large text box on the BLAT
search page (Fig. 5.3), change the Freeze pull-down menu to Dec.
2001, change the Query pull-down menu to DNA and then press
Submit. The server will (very quickly) return the search results; in
this case, a single match of length 636 is found on the forward
strand of chromosome 9 (Fig. 5.4).
To obtain more details on this hit, click the details link, to the
left of the entry. A long web page is returned, with three major
sections: the mRNA sequence (Fig. 5.5, top), the genomic
sequence (Fig. 5.5, middle) and an alignment of the mRNA
sequence against the genomic sequence (see Fig. 5.9 for an exam-
ple). In the alignment in Fig. 5.5, matching bases in the cDNA
and genomic sequences are colored in darker blue and capital-
ized. Gaps are indicated in lower-case black type. Light blue
upper-case bases mark the boundaries of aligned regions on
either side of a gap and are often splice sites.
Returning to the BLAT summary page for this search (Fig. 5.4),
click on browser. This will produce a graphic representation of
where this particular mRNA sequence aligns to the genome
(Fig. 5.6). The track labeled Chromosome Band indicates that the
mRNA maps to 9q34.11. The query sequence itself is represented
on the line labeled Your Sequence from BLAT Search (arrow,
Fig. 5.6). The sequence is shown as being discontinuous: regions
of similarity are shown as vertical lines, gaps are shown as thin
horizontal lines, and the direction of the alignment is indicated
by the arrowheads. The aligned regions of the EST query corre-
spond to the exons of a known gene, shown on the line immedi-
ately below (Known Genes, here RAB9P40). Typing the EST
name, BG334944, directly into a UCSC search box would have
generated a similar result to that shown in Fig. 5.6, but part of the
purpose of this example is to illustrate the use of BLAT.
Approximately halfway down the graphic is a track labeled
Human ESTs That Have Been Spliced. This track is at first shown
in dense mode, with all the ESTs condensed onto a single line. To
see all of the ESTs that align with the genome in this region,
supplement to nature genetics • september 2002
user’s guide
potentially representing differentially spliced transcripts, click
on the track’s label. This will expand this area of the figure so
that each EST occupies a single line (Fig. 5.7). The ESTs are of
varying length, but most contain the same exons as the known
gene and are (presumably) spliced in the same way. Close
inspection indicates that some of the ESTs are missing one or
more exons compared with the known gene. Consider the lines
marked BE798864 and W52533: the former appears to be miss-
ing the fifth exon, whereas the latter is missing the fourth, fifth
and sixth exons.
Any of the ESTs can be examined in more detail by clicking on
that particular line. Here, click on the line for BE798864 (arrow,
Fig. 5.7) to reach the information page for this EST (Fig. 5.8).
The EST is 99.8% identical to the genomic sequence; clicking
anywhere on the hyperlinked line in the section marked
EST/Genomic Alignments returns the actual side-by-side align-
ment (Fig. 5.9). Differences exist at the ends of the EST, but the
sequences are identical in the region surrounding the putative
missing exon.
An alternatively spliced mRNA is more likely to be of biologi-
cal significance when it changes the sequence of the encoded,
wildtype protein. To determine whether EST BE798864 could
encode a protein different from that of the known gene
(RAB9P40), one can simply compare the two sequences directly
against each other using the NCBI’s BLAST 2 Sequences tool.
First, open a new web browser window, because information
from the above search will be needed here; this will prevent hav-
ing to use the browser’s Back and Forward keys excessively and is
a good general rule when using multiple web tools. Then access
the BLAST home page, at http://www.ncbi.nlm.nih.gov/BLAST.
Select BLAST 2 Sequences, under the header labeled Pairwise
BLAST. On this page, the user can simply enter accession num-
bers rather than cutting and pasting sequences into the text
boxes. For the EST, simply enter its accession number
Ensembl also displays database hits that overlap with each
exon in a transcript. These hits may include proteins as well as
ESTs and mRNAs, and may illustrate alternatively spliced
products. The hits are shown as green boxes in the TransView
(Fig.13.5), which can be accessed in a number of ways; for
example, by clicking on the View Evidence box for a transcript
on the GeneView (Fig. 1.10). Another good starting point for
visualizing alternatively spliced transcripts is the NCBI’s
Model Maker (follow the mm link in Fig. 1.2). The Model
Maker displays putative exons from mRNAs, ESTs and gene
predictions that align with the genome. Users can select indi-
vidual exons from these alignments and build a customized
gene model. As the Model Maker displays the nucleotide
sequence of the model along with its three-frame translation,
the effects of adding, modifying or deleting exons can be
quickly evaluated.
33
 user’s guide
(BE798864) into the box marked Enter accession or GI for
Sequence 1. Obtaining the accession number of RAB9P40
requires going back to the graphic shown in Fig. 5.6 and clicking
on the gene’s track. Once this has been done, input the gene’s
accession number (NM_005833) into the box marked Enter
accession or GI for Sequence 2. Make sure that the Program pull-
down is set to blastn (to compare a nucleotide sequence against
another nucleotide sequence, hence the n in blastn) and click the
Align button at the bottom of the page to generate the alignment
(Fig. 5.10). The sequence corresponding to sequence 1 (the EST)
is denoted as the query, whereas the sequence corresponding to
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics
sequence 2 (the known gene) is denoted as the subject. The
known gene’s protein translation is also shown, starting at the
end of the third row of the alignment. Examination of the align-
ment shows that the EST is missing 153 nt (nt 360–512 of the
34
mRNA), which corresponds to the fifth exon that is missing in
BE798864. This gap is in frame, so the EST could encode a
homologous yet shorter protein.
Because of the nature of EST sequencing, ESTs often contain
sequencing errors at a rate much higher than those of the fin-
ished or even draft genomic sequence. It is certainly encouraging
that EST BE798864 aligns well with the genomic sequence and
that its encoded protein could be in the same frame as that pro-
duced from the known gene. In addition, it appears from the
UCSC graphic (Fig. 5.7) that other ESTs in this region, such as
BE779110, are also missing the fifth exon of RAB9P40. All these
predictions must, however, be tested computationally by looking
at the quality of the EST–genomic alignment as shown above.
Final proof of alternative splicing can, of course, only be gener-
ated at the laboratory bench.
supplement to nature genetics • september 2002
 user’s guide
Figure 5.1
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 5.2

supplement to nature genetics • september 2002 35

 user’s guide
Figure 5.3
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 5.4

36 supplement to nature genetics • september 2002

 user’s guide
Figure 5.5
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 5.6

supplement to nature genetics • september 2002 37

 user’s guide
Figure 5.7
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 5.8

38 supplement to nature genetics • september 2002

 user’s guide
Figure 5.9
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 5.10

supplement to nature genetics • september 2002 39

 user’s guide
Question 6
How would one retrieve the sequence of a gene, along with all anno-
tated exons and introns, as well as a certain number of flanking bases for
use in primer design?
doi:10.1038/ng971
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics
This type of search can be initiated at the UCSC Genome
Browser home page, located at http://genome.ucsc.edu. Select
Human from the pull-down menu labeled Organism, and then
click on Browser. This brings the user to the Human Genome
Browser Gateway, from which a number of text- or position-
based searches can be performed on current or older versions of
the genome assembly. In this case, select the Dec. 2001 assem-
bly, type the name of the gene of interest (PTPN1) into the posi-
tion box, and then click Submit. The Browser returns all genes
starting with the characters ‘PTPN1’ (Fig. 6.1). The gene of
interest here is the one called PTPN1; click on the hyperlinked
PTPN1 (arrow, Fig. 6.1) to view the genomic context of this
gene (Fig. 6.2).
The text box at the top of Fig. 6.2 gives the absolute base pair
position of this gene (chromosome 20, positions
48929540–49003636) and indicates that the gene spans 74 kb.
The track labeled Chromosome Bands shows that PTPN1 is
located at 20q13.13. Finally, the track marked Known Genes
shows that the gene is on the forward strand, as the arrows on
that track are pointing to the right. The exons within this gene
are indicated by the vertical lines in the Known Genes track.
One way to obtain sequence upstream of a gene is described in
Question 7. Here we explain how to retrieve flanking sequence
on both sides of a gene. To retrieve an adequate amount of
sequence with which to design primers, one can increase the size
of the region displayed by changing the position numbers within
the position box at the top of the figure. To add an additional
1,000 nt at the 5′ end and an additional 200 nt at the 3′ end, for
example, change the text in the position box to ‘chr20:4892854-
49003836’ and click Jump. This now redraws the graphic with the
new boundaries.
To obtain the actual sequence within the region, click on the
DNA link in the blue bar at the top of the page. This produces a
new page, entitled Get DNA in Window (Fig. 6.3). Click the but-
40
ton next to extended case/color options and then click Submit. By
selecting this option, the user can highlight features in the
sequence by changing the format (case, underline, bold, italic)
and/or color (red, green, blue) of the text. Colors can be varied in
darkness and mixed together by changing the values in the boxes
under Red, Green and Blue to any number between 0 and 255;
examples of how to specify in RGB (red-green-blue) format color
are given below the table. At this point, check the Toggle Case box
in the Known Genes row, change the red saturation to 255 and
leave the other color values set at zero (Fig. 6.4). Once the user
clicks Submit, a new page is presented with the entire length of
the sequence specified above (chr20:48928540-49003836) and
the exons within this range are shown in red in capital letters
(Fig. 6.5). This genomic sequence can now be saved and
imported into a primer design or sequence assembly package for
further analysis.
The Extended DNA Case/Color Options page can be used to
combine and differentiate between genomic tracks. For exam-
ple, return to the Options page, leave the Known Genes row as
before but now also check the Underline square in the Mouse
Blat row of the table. Clicking Submit produces a page on which
the human exons still appear in red capital letters, but hits from
the mouse sequence are now shown as underlined text (Fig. 6.6).
In this section of the gene, the conserved mouse sequence over-
laps with the exons.
One way to retrieve sequence for a defined chromosomal
region at the NCBI is with the seq link on the MapViewer, vis-
ible when the Gene_Seq map is the master (Fig. 1.2). At
Ensembl, export genomic nucleotide sequence with the
Export→FASTA link in any ContigView window (Fig. 1.14,
center yellow bar).
supplement to nature genetics • september 2002
 user’s guide
Figure 6.1
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 6.2

supplement to nature genetics • september 2002 41

 user’s guide
Figure 6.3
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 6.4

42 supplement to nature genetics • september 2002

 user’s guide
Figure 6.5
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 6.6

supplement to nature genetics • september 2002 43

 user’s guide
Question 7
How would an investigator easily find compiled information describing
the structure of a gene of interest? Is it possible to obtain the sequence
of any putative promoter regions?
doi:10.1038/ng972
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics
One place to initiate this search is at UCSC’s Genome Browser, at
http://genome.ucsc.edu. For purposes of this example, consider
the gene encoding pendrin (PDS), a protein associated with
developmental abnormalities of the cochlea, sensorineural hear-
ing loss and diffuse thyroid enlargement (goiter).
From the UCSC home page, choose Human from the pull-
down Organism list, and click on Browser. The user is now at the
Human Genome Browser Gateway. The search in this case is sim-
ple: select Dec. 2001 from the assembly pull-down menu, type
pendrin into the position box, and then click Submit. The
returned results indicate one known gene and two mRNA
sequences; click on the accession number of the mRNA sequence
AF030880 to continue. The user will now be presented with a
graphic overview of the region containing this mRNA. To gain a
better perspective of the region, click on the 1.5× button next to
zoom out. Finally, click the reset all button on the middle of the
page to reset the tracks to their default settings.
Carrying out these steps will produce an output similar to that
shown in Fig. 7.1. For the purpose of this question, however, the
default settings are not ideal. Using the Track Controls at the bot-
tom of the figure, and following the example in Fig. 7.2, set some
tracks to hide mode (not shown), others to dense (all data con-
densed onto one line) and some to full (a separate line for each
feature, up to 300). Before considering the actual data within
these tracks, a brief discussion of the content and representation
of these tracks is warranted. Many were provided to UCSC by
outside individuals. Further information on the gene prediction
methods briefly discussed below can be found elsewhere15.
The general convention for the Known Genes and predicted
gene tracks (Fig. 7.1) is that each coding exon is shown as a tall,
vertical bar or block. 5′ and 3′ untranslated regions are shown as
shorter vertical bars or blocks.
Connecting introns are shown as very thin lines. The direction
of transcription is indicated by the arrows along that thin line.
Known Genes are taken from mRNA reference sequences
within LocusLink10. These reference sequences have been aligned
against the genome using BLAT.
The Acembly Gene Predictions With Alt-splicing track is derived
from the alignment of human mRNA and EST sequence data
against the genome, using the program Acembly. This program
attempts to find the best alignment of each mRNA against the
genome and considers alternative splice models. If more than
one gene model with statistical significance can be produced,
each of these is shown in the display. Additional information on
Acembly can be found on the NCBI web site at
http://www.ncbi.nih.gov/IEB/Research/Acembly/.
The Ensembl Gene Predictions track7 is provided by Ensembl.
The Ensembl genes are predicted by a range of methods, includ-
ing homology to known mRNAs and proteins, ab initio gene pre-
diction using GENSCAN and gene prediction HMMs.
The Fgenesh++ Gene Predictions come from a method that pre-
dicts internal exons by looking for structural features such as
donor and acceptor splice sites, putative coding regions and
44
intronic regions both 5′ and 3′ to a putative exon using a dynamic
programming algorithm; the method also takes into account
protein similarity data16.
The Genscan Gene Predictions derive from a method called
GENSCAN, through which introns, exons, promoter sites and
poly(A) signals can be identified. Here, the method does not
expect the query sequence to represent one and only one gene, so
it can make accurate predictions for either partial genes or multi-
ple genes separated by intergenic DNA11.
The Human mRNAs from Genbank track shows alignments
between human mRNAs in GenBank and the genome sequence.
The Spliced ESTs and Human EST tracks show the alignment of
ESTs from GenBank against the genome. Because ESTs usually
represent fragments of transcribed genes, there is high likelihood
that an EST corresponds to an exonic region.
Finally, the Repeating Elements by RepeatMasker track shows,
as its name would suggest, repetitive elements such as short and
long interspersed nuclear elements (SINEs and LINEs), long ter-
minal repeats (LTRs) and low-complexity regions (http://repeat-
masker.genome.washington.edu/cgi-bin/RepeatMasker). It is
customary to remove or ‘mask’ these elements before applying a
gene prediction method to a nucleotide sequence.
Returning to the example shown in Fig. 7.2, notice that most of
the tracks return a nearly identical gene prediction; as a rule,
exons predicted by multiple methods increase the likelihood that
the prediction is actually correct and does not represent a ‘false
positive’. Most of the methods show a 3′ untranslated region, indi-
cated by the heavy, shorter block at the left of the predictions. The
Acembly track shows three possible alternative splices in addition
to the full-length product shown in the third line of that section, a
prediction that agrees with those shown in most of the other
tracks. The Genscan track extends off to both the right and the
left: GENSCAN can be used to predict multiple genes, and this
display implies that the method has been applied in this fashion.
Although these graphical overviews are useful, the investigator
will more often than not want the actual sequence corresponding
to these blocks. For this example, the Fgenesh++ prediction will
be used as the basis for obtaining raw sequence data, but the steps
will be identical regardless of which track is chosen. Click on the
track labeled Fgenesh++ Gene Predictions to go to a summary
page describing the prediction (Fig. 7.3). The region has sequence
similarity to the pendrin gene (which was already known at the
beginning of the example). The size and the beginning- and end-
points of the prediction are given, and it is indicated that the pre-
diction lies on the minus strand; this was also indicated in Fig. 7.2
by the left-pointing arrows in the intronic regions. To obtain the
sequence, click on Genomic Sequence. The user will be taken to a
The NCBI also provides gene predictions, computed using
the program GenomeScan17. These models are shown on the
GenomeScan and Gene_Seq maps.
supplement to nature genetics • september 2002

query page entitled Get Genomic Sequence Near Gene, from which
the transcript, coding region, promoter, or both the transcript
and promoter can be obtained (Fig. 7.4). For each of the options,
the sequence is returned in FASTA format, with the nucleotide
coordinates being given in the definition line.
Transcript returns the sequence of the entire transcript, with
exons shown in upper-case letters.
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics
supplement to nature genetics • september 2002
user’s guide
Coding Region Only returns just the coding region, with exons
shown in upper-case letters.
Transcript + Promoter appends the promoter sequence to the 5′
end of the sequence that the user would have obtained by using
the Transcript option, with exons shown in upper-case letters.
The length of the promoter can be indicated in the text box.
Promoter returns just the promoter region, as shown in Fig. 7.5.
45
 user’s guide
Figure 7.1
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 7.2

46 supplement to nature genetics • september 2002

 user’s guide
Figure 7.3
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 7.4

supplement to nature genetics • september 2002 47

 user’s guide
Figure 7.5
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

48 supplement to nature genetics • september 2002


Question 8
How can one find all the members of a human gene family?
doi:10.1038/ng973
The HUGO Gene Nomenclature Committee (http://www.gene.
ucl.ac.uk/nomenclature/) has been working to develop a unique
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics
symbol, as well as a longer and more descriptive name, for each
human gene. Thus, members of many gene families, previously
cloned in different laboratories and known by a variety of terms,
now share a common gene symbol. A text search in any of the
genome browsers will often return links to all named members of
a gene family that have been mapped to the genome. Whereas
Ensembl and UCSC currently return lists of the genes, the NCBI
presents both a list and a graphical overview.
Go to the NCBI home page at http://www.ncbi.nlm.nih.gov/
and click on the Human map viewer link on the right side to
access the Map Viewer search page. Enter the term
‘ADAM*[sym]’ in the text query box. The asterisk, or wild card,
will match any character, whereas the term [sym] limits the
search to items with ADAM as their gene symbol. Other
advanced search options are available by clicking the Advanced
Search box or by reading the online documentation. The search
returns 41 hits, which include members of the ADAM family as
well as other related families whose names start with the term
‘ADAM’, such as ADAMTS and ADAMDEC. To limit the search
to ADAM genes only, eliminate the undesired gene symbols with
the Boolean NOT term, using the query ADAM*[sym] NOT
ADAMTS*[sym] NOT ADAMDEC1*[sym]. The graphic at the
top of the returned page shows the location of each gene with a
red tick mark (Fig. 8.1). It is immediately clear that the 19
mapped ADAM genes are distributed among 11 chromosomes,
and that some, such as those at the tips of the q arms of chromo-
somes 10 and 14, are close together. The list at the bottom of the
page presents links to the 19 genes.
Another way to search for homologous genes in the genome is
through a basic local alignment search tool (BLAST) search at
the NCBI or Ensembl. BLAT searches at UCSC are not as sensi-
tive as BLAST searches and may not find as many homologous
genes. In this example, all genomic sequences homologous to the
ADAM2 protein will be found using the Ensembl BLAST inter-
face. From the Ensembl Human home page at
http://www.ensembl.org/Homo_sapiens/, click on the link to
supplement to nature genetics • september 2002
user’s guide
BLAST. Paste the sequence of the ADAM2 protein (GenBank
accession NP_001455.2) into the query box (having obtained the
protein sequence from the NCBI’s Entrez database by following
the steps in Question 5). Set the database to Homo sapiens,
genomic sequence to search the Ensembl genome assembly, and
choose TBLASTN as the executable (Fig. 8.2). Use the default
parameters for the remaining settings. When done, click Search.
The returned page will contain a retrieval ID (Fig. 8.3), which,
when the search is finished, will link to the search results page
(Fig. 8.4).
The top of the results page shows a graphical overview of the
locations of hits. These hits may be to the entire protein or just to
a single domain. The hits are colored by BLAST score, red being
most similar, blue least similar and green intermediate. Some of
the hits, like the pairs on the q arms of chromosomes 10 and 14,
lie in positions similar to those of ADAMs mapped by the NCBI
(Fig. 8.1), but others, such as those on chromosomes 12 and Y,
are unique to the BLAST search. These unique hits may represent
real members of the ADAM family that have not yet been named
and would therefore not show up in a text-based search. Alterna-
tively, they may be unnamed pseudogenes or nonsignificant
BLAST hits. One gene on chromosome 1 is found in the text-
based search at the NCBI but not in the BLAST search at
Ensembl. The similarity between this gene and ADAM2 is not
high enough for it to appear in the BLAST search using the
default Ensembl parameters.
Clicking on an arrow next to one of the hits shown in Figure 8.4
activates a pop-up menu that gives the details of the BLAST
report and provides links to the BLAST alignment and the
ContigView (Figs 8.5 and 8.6, respectively, for the hit on chromo-
some 12). The hit on chromosome 12 contains a stop codon and
is probably an intronless pseudogene. The bottom of the results
page (Fig. 8.4) shows a summary of the BLAST hits. Clicking on a
hit links to the BLAST alignment (Fig. 8.5). A link in the middle
of the results page (Fig. 8.4) provides the entire BLAST report in
standard format. Clicking on a hit in the BLAST report retrieves
the ContigView for the region around the hit (similar to what is
shown in Fig. 8.6).
49
 user’s guide
Figure 8.1
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 8.2

50 supplement to nature genetics • september 2002

 user’s guide
Figure 8.3
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 8.4

supplement to nature genetics • september 2002 51

 user’s guide
Figure 8.5
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 8.6

52 supplement to nature genetics • september 2002

 user’s guide

Question 9
Are there ways to customize displays and designate preferences? Can
tracks or features be added to displays by users on the basis of their
own research?
doi:10.1038/ng974
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

In this example, the UCSC browser will be used to view particu-
lar tracks. Start at the UCSC home page (http://genome.ucsc.
edu), click on Browser in the blue sidebar on the left-hand side of
the page, and set the Genome Browser Gateway to a region of
interest. For example, one could set the genome to Human and
the assembly to Dec. 2001, type chr22:38496887-39496866 into
the position box, and click Submit to display a representative
region of the December 2001 assembly of human chromosome
22. A number of tracks are already displayed in dense format
(Fig. 9.1). Below the graphic showing the specified region are
pull-down menus that allow the user to change the appearance of
the graphic, under the heading Track Controls (Fig. 9.2). There
are three options in each of these pull-down menus:
• Hide, which allows the user to eliminate that particular track
from the display.
• Dense, which displays all annotations or features for that
track on a single line.
• Full, which displays each annotation or feature for that track
on a separate line; this is the ‘exploded view’ that is illustrated
in a number of the questions in this guide.
Once the desired selections have been made, the user clicks on
the refresh button to redraw the graphic. Further customization
of individual tracks can be achieved by clicking on the track
name in the Track Controls section of the browser. The user can,
for example, customize the EST track controls to color red all
ESTs from a certain library that contain a particular keyword in
their GenBank entry or to eliminate all such ESTs from the dis-
play. The browser retains these selections for all subsequent ses-
sions; the default settings can be restored by clicking on the reset
all button.
One of the attractive features of the UCSC system is that users
can add their own annotations, features or tracks to their local
displays. These changes are not written or saved in any way to the
original data held at UCSC. To customize the display, the user
returns to the Human Genome Browser Gateway page and scrolls
down to the Add Your Own Tracks section. Here, the user is pre-
sented with a large text box into which properly formatted text
can be typed or pasted. Alternatively, the specifications can be in
a text file, which the user can select by using the Browse button
above the large text box. As another option, if the text file is
posted on the user’s local web page, the user can share the custom
track of annotations with other colleagues simply by telling them
the URL of the file. Colleagues can then view the custom annota-
tion by starting the UCSC browser and entering this URL into
the large text box.
For the purposes of this example, enter the following text
file into the entry field (Fig. 9.3) and click Submit at the top of
the page:

browser position chr22:38496887-39496866

browser hide cytoBand
browser hide stsMap
browser hide gap
browser hide clonePos
browser full refGene
browser dense mrna
track name="scale" description="our peak"
chr22 38996887 38996888 peak
track name="Microsatellites" description="Microsatellites" color=0,128,0
chr22 38627059 38627060 D22S276
chr22 39005417 39005418 D22S307
track name="Genotyped SNPs" description="Genotyped SNPs" color=0,0,255
chr22 38518342 38518343 ss146131
chr22 38705963 38705964 ss2941443
chr22 38884157 38884158 ss141110
chr22 39171390 39171391 ss22916
chr22 39438769 39438770 ss1479794
track name="Upcoming SNPs" description="Upcoming SNPs" color=0,128,192
chr22 38615712 38615713 ss86855
chr22 38804838 38804839 ss85533
chr22 39077895 39077896 ss141190
chr22 39305065 39305066 ss137027

supplement to nature genetics • september 2002 53

 user’s guide
This is but one example using only some of many options to
At Ensembl, the user can customize the Detailed View section the Add Your Own Tracks feature. A full description, information
of the ContigView, adding annotations and changing colors, on input format and additional examples are available at
by selecting options under the Features link (Fig. 1.14, center http://genome.ucsc.edu/goldenPath/help/customTrack.html.
yellow bar). Users can visualize their own custom data on
Ensembl’s ContigView displays and even share that data with Table 9.1 Symbolic Names for UCSC Browser Tracks
other users by following the instructions under the DAS Track Symbolic name
Sources link (Fig. 1.14, center yellow bar). At the NCBI, Map
Acembly Genes acembly
Viewer displays are changed in the Maps & Options window Assembly gold
(Fig. 3.9). BAC End Pairs bacEndPairs
Base position ruler
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Chromosome Band cytoBank

Coverage clonePos
The browser will ignore any entries in the position box and CpG Islands cpgIsland
look only to the file pasted in the Add Your Own Tracks field. The Duplications genomicDups
results of this customized display are shown in Fig. 9.4. Ensembl Genes ensGene
Lines that begin with the word ‘browser’ control the overall Exofish Ecores exoFish
Fgenesh++ Genes softberryGene
browser display. Lines beginning with ‘track’ create new tracks. Fish Blat blatFish
Lines following track lines provide positional information for FISH Clones fishClones
each item to be displayed on that track. Therefore: Gap gap
• the first line of the above format sets the browser to position GC Percent gcPercent
Geneid Genes geneid
38496887–39496866 on chromosome 22. Genscan Genes genscan
• the next six ‘browser’ lines change the overall browser display GNF Ratio affyRatio
for the Chromosome Band, STS Markers, Gap, Coverage, Human Blat blatHuman
Known Genes and Human mRNAs tracks. The formatted text Human ESTs est
must contain the symbolic name for each track rather than Human mRNAs mrna
Known Genes refGene
the name listed on the web page display. Symbolic names Map Contigs ctgPos
used by the UCSC browser are listed in Table 9.1. Compared Mouse Blat blatMouse
with the default settings (Fig. 9.1), the Chromosome Band, Mouse ESTs est
STS Markers, Gap and Coverage tracks have all been hidden, Mouse mRNA mrna
Mouse Synteny mouseSyn
and Human mRNAs is dense rather than full (Fig. 9.4). NCI60 nci60
• the remaining lines instruct the browser to create four new Nonhuman EST xenoEst
tracks named scale, Microsatellites, Genotyped SNPs and Nonhuman mRNA xenoMrna
Upcoming SNPs, respectively. Names are listed on the left side Nonmouse mRNA xenoMrna
Overlap SNPs snpNih
of the browser display. The lines beginning with the word Random SNPs snpTsc
‘track’ name the tracks, as listed above, and also set the RepeatMasker rmsk
descriptions (our peak, Microsatellites, Genotyped SNPs and Rosetta rosetta
Upcoming SNPs) and colors [default (black), green, blue, and Sanger 22 sanger22
Simple Repeats simpleRepeat
light blue] to be used to display those tracks (Fig. 9.4). The Spliced ESTs intronEst
descriptions appear as a center label in the browser and colors STS Markers stsMap
are determined by the three RGB values provided. All lines Tigr Gene Index tigrGeneIndex
following a ‘track’ line provide position information for the UniGene uniGene
tick marks corresponding to the individual items. For exam-
ple, ‘peak’ is displayed at position 38996887–38996888 on
chromosome 22.

54 supplement to nature genetics • september 2002

 user’s guide
Figure 9.1
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 9.2

supplement to nature genetics • september 2002 55

 user’s guide
Figure 9.3
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 9.4

56 supplement to nature genetics • september 2002


Question 10
For a given protein, how can one determine whether it contains any
functional domains of interest? What other proteins contain the same
functional domains as this protein? How can one determine whether
there is a similarity to other proteins, not only at the sequence level, but
also at the structural level?
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics
doi:10.1038/ng975
To demonstrate how to find functional domains within a protein,
the human testis-determining factor TDF, also known as the sex-
determining protein SRY, will be used as an example.
Although the search could be commenced from the Entrez
search box on the NCBI home page, a better way to perform the
initial search is from LocusLink10. One of the advantages of
using LocusLink lies in its standardization of gene and protein
names with appropriate cross-referencing, making it more
likely that the correct protein will be found on the first attempt.
From the NCBI home page at http://www.ncbi.nlm.nih.gov/,
choose LocusLink from the pull-down menu in the upper left
corner, type the gene name, ‘TDF’, into the query box, and click
Go. Four loci are returned (Fig. 10.1). The first column gives
the Locus ID, which is a stable identifier associated with that
gene locus. Clicking on the LocusID produces a LocusLink
report view; more detailed information on the report view can
be found in the LocusLink Help feature and in the literature15.
The second column, marked Org, gives a shorthand version of
the organism name. Here, there is one entry from Drosophila
(Dm), one from mouse (Mm), one from human (Hs) and one
from rat (Rn). A series of alphabet blocks shown to the right of
each entry provide jumping-off points to other database
resources. The locus of interest here is the third entry in the list,
because that is the one for the human form of TDF/SRY.
To find additional information on the protein, click on the sec-
ond P (in green) on that line. This takes the user to the protein
entries corresponding to that particular LocusLink entry
(Fig. 10.2). At this point, the user can click on any of the hyper-
links to look at the raw database information available on any
of the proteins listed.
Consider the first entry in the list, an NCBI Reference Protein
sequence with accession number NP_003131. To the right of the
accession number is a series of hyperlinks. Clicking on the link
labeled BLink will take the user to the BLink page for the protein
of interest (Fig. 10.3). BLink stands for ‘BLAST Link’ and pro-
vides the graphical results of pre-computed BLAST searches that
have been performed not just for this protein sequence, but for
every protein sequence within the Entrez Proteins data domain.
The pre-computed BLAST results for TDF/SRY are shown in the
section beginning with the label ‘204 aa’. Across the top are a
number of buttons that allow the user to ask a series of questions
regarding their protein of interest. As the object of this question
is to find the protein domains present within the TDF/SRY pro-
tein, the user can click on CDD-Search (Conserved Domain
Database Search18). Doing this will produce a graphical overview
of any domains present within the protein, as well as a sequence
alignment of those domains with the query sequence (Fig. 10.4).
In this case, one functional domain is found: an HMG box,
which is a DNA-binding domain found in many nuclear pro-
teins. The domain was found in both of the databases comprising
supplement to nature genetics • september 2002
user’s guide
CDD (Pfam and SMART), as can be seen by looking at the acces-
sion numbers in the hit list.
To determine which other proteins contain this same HMG-
box domain, click on the box labeled Show, right under the
graphical view near the top of the page. This will invoke the
domain architecture retrieval tool (DART). DART shows func-
tional domains within a protein and, more importantly, other
proteins with a similar domain architecture (Fig. 10.5). The
query (the HMG-box) is shown at the top of the page in red.
Every other protein in the NCBI’s non-redundant sequence data-
base having that same domain is then shown below the query,
with the HMG box again colored red. Other domains within the
found proteins are also shown, in various colors and shapes, with
a key appearing at the bottom of the web page. Clicking on any of
the links to the left would provide additional information about
these new proteins.
Although a protein domain has now been identified within the
query protein, no in-depth information has yet been provided
about the function of that domain. Whereas a circuitous path
could be followed from the DART page to find this information,
an easier method is to use another web-based resource, called
InterPro. InterPro is an integrated resource for information
about protein families, domains and functional sites, bringing
together information from a number of protein domain-based
resources, such as PROSITE, PRINTS, Pfam and ProDom19. The
InterPro Simple Search engine can be accessed from the InterPro
home page, at http://www.ebi.ac.uk/interpro. Clicking on Text
Search, on the left, brings the user to the search page; for this
search, type ‘HMG Box’ into the text box and hit Search. Three
hits are returned (Fig. 10.6). For purposes of this example, follow
the link from the first hit, for high mobility group proteins HMG1
and HMG2 (IPR000135). The resulting InterPro summary page
(Fig. 10.7) provides information on the function, intracellular
location and, most importantly, metabolic role of this particular
protein within the cell, in an executive summary format. Refer-
ences are provided at the bottom of the web page for users who
wish for more in-depth information about the domain. Users
can also retrieve all of the full-length sequences containing the
domain; the reader is referred to the InterPro documentation for
more details.
The final part of this question asks whether similarity to the
query protein can be found at the structural as well as the
sequence level. Answering this question requires a new search
against NCBI Structures. From the NCBI home page, change the
pull-down menu in the query box at the top of the page to Struc-
At Ensembl, the GeneView links directly to the InterPro
domain(s) found in the protein (Fig. 1.9).
57
 user’s guide
ture, type ‘SRY’ in the box and hit Go. Four three-dimensional
structures are returned, one of which is 1HRY, the structure of
the human SRY–DNA complex solved by NMR. Clicking on the
1HRY hyperlink takes the user to the Structure Summary page
for 1HRY. The summary links to more detailed information
about chain A, the protein component of the structure, chain B,
the nucleotide component of the structure, and the conserved
domain (CD) in the protein, obtained through a CDD search.
Click on the chain A graphic to get a list of proteins whose known
structures have, using a method called VAST, been deemed simi-
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics
58
lar to that of the original SRY protein; more information on the
method and on interpreting the data within the tables can be
found elsewhere15. Here, the SRY protein is shown to have some
structural similarity to a fasciculin 2–mouse acetylcholinesterase
complex, a protein named V-1 Nef, a heat-shock protein of 70 kD
and a myosin motor-domain complex (Fig. 10.8). The VAST pro-
gram quite often reveals similarities between proteins that are
not evident from simple BLAST or FASTA searches, so readers
are encouraged to employ this and similar tools when trying to
answer questions related to protein families.
supplement to nature genetics • september 2002
 user’s guide
Figure 10.1
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 10.2

supplement to nature genetics • september 2002 59

 user’s guide
Figure 10.3
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 10.4

60 supplement to nature genetics • september 2002

 user’s guide
Figure 10.5
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 10.6

supplement to nature genetics • september 2002 61

 user’s guide
Figure 10.7
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 10.8

62 supplement to nature genetics • september 2002


Question 11
An investigator has identified and cloned a human gene, but no corre-
sponding mouse ortholog has yet been identified. How can a mouse
genomic sequence with similarity to the human gene sequence be
retrieved?
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics
doi:10.1038/ng976
For purposes of this example, assume that the user does not
already have the human sequence of interest to hand. The first
step will be to locate the human gene of interest using the UCSC
Genome Browser. Begin by pointing to the UCSC Genome
Browser home page, at http://genome.ucsc.edu. Select Human
from the Organism pull-down menu and then click on Browser;
both are located on the blue navigation bar at the left side of the
page. This will take the user to the Human Genome Browser
Gateway. Select the Dec. 2001 version of the UCSC genome
assembly, type the gene symbol ‘AGPS’ into the position box, and
then click Submit. On the resulting page, follow the link for
AGPS in the Known Genes section.
The result of the search on AGPS is shown in Fig. 11.1. In the
main figure are a series of ‘tracks’, which are labeled along the
left-hand side. The Known Gene track is for AGPS, correspond-
ing to the query. Clicking on AGPS returns a summary of infor-
mation on that gene, including the full name of the protein
product (alkylglycerone phosphate synthase precursor), a link to
the GeneCards database at the Weizmann Institute20 and links to
the translated protein, mRNA and genomic sequences. Focus
now on the track labeled Mouse Translated Blat Alignments.
What is shown in this track are the results of aligning the
November 2001 version of the mouse genome assembly with the
human genome using the program BLAT8 in its translated pro-
tein mode. More details about the BLAT algorithm and about
how the mouse BLAT track is automatically generated can be
found by clicking on the Mouse Blat hyperlink found below the
main graphical display.
Click anywhere within the Mouse Blat track to expand the sin-
gle BLAT track so that it now shows each individual mouse
sequence that aligns with human sequence in the region of inter-
supplement to nature genetics • september 2002
user’s guide
est (Fig. 11.2). Especially in a translated mode, mouse and
human gene sequences are usually more similar in exons than in
introns. Look carefully at the two alignments that derive from a
mouse sequence called chr3 81178k (Fig. 11.2, arrow). On the
Mouse Blat track, the brown vertical lines represent alignments
and the horizontal lines are gaps. These alignments correspond
to the blue vertical lines indicating the exons of AGPS on the
Known Genes track.
To see the kind of information available for a translated BLAT
alignment, click on the mouse genomic sequence labeled chr3
81178k. The resulting page (Fig. 11.3) provides the details of the
alignment of the trace with the human genome assembly. This
mouse genomic sequence is 607 nt in length and aligns with the
human sequence in eight blocks. Within the blocks, the mouse
and human sequences are 78% identical. To view the alignment
itself, click on the View details of parts of alignment. . . link. On
the resulting page (Fig. 11.4), the mouse sequence is shown on
top, with the region of alignment in blue. The human genomic
sequence is shown next, and a side-by-side alignment of the
human and mouse sequences is at the bottom of the web page
(not shown).
The NCBI’s UniGene_Mouse map shows alignments of
mouse mRNA and EST sequences with the human genome.
Add this map using Maps & Options (Fig. 3.9). The easiest
way to find the mouse ortholog of a human gene is probably
to use Ensembl’s precomputed Homology Matches. These
matches, where available, link directly from a human gene to
a putative mouse homolog (Fig. 1.9).
63
 user’s guide
Figure 11.1
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 11.2

64 supplement to nature genetics • september 2002

 user’s guide
Figure 11.3
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 11.4

supplement to nature genetics • september 2002 65

 user’s guide
Question 12
How does a user find characterized mouse mutants corresponding to
human genes?
doi:10.1038/ng977
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics
The NCBI provides a set of maps that show chromosomal
regions homologous between mouse and human. This resource
can be accessed directly at http://www.ncbi.nlm.nih.gov/Homol-
ogy/. For this example using a known, mapped human gene,
however, it is easier to start the search from the LocusLink entry
for tyrosinase. The LocusLink10 query page can be found at
http://www.ncbi.nlm.nih.gov/LocusLink/. Select Human from
the Organism pull-down menu, enter ‘tyrosinase’ into the Query
box and click Go. To view the entry for tyrosinase (TYR), click on
its LocusLink ID, 7299.
On the resulting page (Fig. 12.1), links to the mouse homology
maps are in the section of the LocusLink summary page marked
Relationships. In this case, there are four maps available for TYR
showing mouse alignments: NCBI vs MGD aligns the current
NCBI assembly of the human genome with the MGD (Mouse
Genome Database21, at The Jackson Laboratory) genetic map,
UCSC vs. MGD aligns the current UCSC genome assembly with
the MGD genetic map, NCBI vs. EST-based RH Map aligns the
NCBI assembly with the Whitehead–MRC RH map, and UCSC
vs. Hudson et al. aligns the 7 October 2001 UCSC assembly with
the Whitehead–MRC RH Map22. The Hs and Mm links adjacent
to each map name show the mouse–human homology map with
the master chromosome as human or mouse, respectively. Click
on the Hs link next to the NCBI vs. MGD map.
The resulting mouse–human map shows the mouse genes that
are the likely orthologs of human genes on human chromosome
11 (Fig. 12.2). Depending on the browser being used, one may
have to click the View as text box to obtain the output; the result-
ing output will appear in text format, slightly different from what
is shown in Fig. 12.2. Chromosomal locations of the mouse genes
are shown, where known. The green circles link to the UniSTS
entry for each locus; those on the left link to the human UniSTS
entry, whereas those on the right link to the mouse UniSTS entry.
The cytogenetic positions are hyperlinked to either the human or
mouse Map Viewer, as appropriate. Gene symbols are linked to
LocusLink10. The tyrosinase gene, highlighted in pink, maps to
mouse chromosome 7 at 44 cM, a piece of information that will
be needed in the next step.
The mouse models themselves are described at the Mouse
Genome Informatics site at the Jackson Laboratory. Go to
the Mouse Genome Informatics home page, at http://www.
informatics.jax.org/, and use the Query Forms pull-down menu
66
to select Linkage Maps. On the resulting page, customize the
search to find the region around the mouse gene Tyr. Under
Chromosome, set the number to 7; then set the chromosomal
region to between 40 and 48 cM.
Many of the uncloned mouse mutants are not mapped in
high-resolution crosses, and many are carried out with a small
number of mice relative to another easy-to-score phenotype for
another mouse mutant that maps to the same chromosome. It is
thus necessary to be lenient in looking for potential uncloned
mouse mutants (±4 cM relative to the location). In this case, as
the NCBI data tells us that the gene is at 44 cM, the region from
40 to 48 cM should be searched.
Further down the page (Fig. 12.3), under Markers, set Include
DNA segments to No to reduce the number of markers shown. Do
include syntenic markers, which are DNA markers and mutant
alleles linked to chromosome 7 that have not been finely mapped
but that may be associated with a phenotype of interest relative to
TYR. Under Comparative Maps, Show homologs from species,
choose human (Homo sapiens). Select Show all markers. Use the
default setting for all other options, and hit Retrieve.
The gene Tyr is found on page 2 of the output, at 44 cm
(Fig. 12.4). The mouse chromosome is shown schematically on
the left and expands as one moves to the right. In the rightmost
columns are the names of the mouse markers in a particular
region in blue and, if there is a corresponding human ortholog,
the name of that ortholog in black. Some of the displayed
mouse markers are genes, some are STSs, some are recessive
mutants (all small letters) and some are dominant alleles (ini-
tial capital letter). At the bottom of the page are syntenic mark-
ers, those which have been mapped to chromosome 7 but not to
an exact position.
Clicking the blue Tyr link at 44 cm opens up a summary of the
Genes, Markers, and Phenotypes for that gene (Fig. 12.5). Of par-
ticular interest in this case are the phenotypic alleles. There are 99
mouse strains with mutations in the Tyr gene.
Users can also view chromosomal regions that are homolo-
gous between mouse and human by using Ensembl’s Syn-
tenyView, available from the ContigView by clicking the Jump
to→syntenyview link (Fig. 1.14, center yellow bar).
supplement to nature genetics • september 2002
 user’s guide
Figure 12.1
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 12.2

supplement to nature genetics • september 2002 67

 user’s guide
Figure 12.3
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 12.4

68 supplement to nature genetics • september 2002

 user’s guide
Figure 12.5
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

supplement to nature genetics • september 2002 69

 erratum

A User’s Guide to the Human Genome

T G Wolfsberg, K A Wetterstrand, M S Guyer, F S Collins & A D Baxevanis
Nature Genet. 32, 1–79 (2002).
doi:10.1038/ng977
Owing to a production error, Figure 11.2 was inadvertently inserted in place of Figure 12.2 in both the print and online PDF versions of
the User’s Guide. The screen pictured in the full text version online is the correct Figure 12.2.

nature genetics
 user’s guide
Question 13
A user has identified an interesting phenotype in a mouse model and
has been able to narrow down the critical region for the responsible
gene to approximately 0.5 cM. How does one find the mouse genes in
this region?
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics
doi:10.1038/ng978
Ensembl provides a mouse genome browser, similar to the one
available for humans. It is being updated with the latest mouse
genome sequence assemblies and, at the time of writing, dis-
plays the MGSC version 3 assembly of the mouse genome, with
sequence data from February 2002. The sequence is estimated
to cover 96% of mouse euchromatic DNA, and Ensembl has
predicted that it contains over 22,000 genes. Start at the
Ensembl mouse home page, at http://www.ensembl.org/
Mus_musculus/. Choose Marker from the pull-down menu,
type the marker name ‘RH114718’ in the adjacent box, and
press Lookup. Click either of the resulting links to view more
details about this radiation hybrid marker. RH114718 has been
mapped to a single position on chromosome 19 and is also
known as MGI:102447, MTH1904 and D19MIT109 (Fig. 13.1).
Click on the chromosomal position to view the genomic con-
text of the marker (Fig. 13.2).
The Overview section of Fig. 13.2 shows a region of 1 Mb of
chromosome 19 centered around the marker, labeled
D19MIT109 in this view. More than 30 mouse genes are pre-
dicted in this region, some already known and some new. The
Detailed View at the bottom of the page is a zoomed-in display of
the region around the marker. To get a better view of the genes
and transcripts in this region, zoom out on the bottom view by
clicking on the longest bar in the zoom control (closest to the
minus sign). The Detailed View will now show the same region of
chromosome 19 as the overview, but with many additional fea-
tures (Fig. 13.3). The splice patterns of the genes and gene pre-
dictions are shown, as are regions of homology between the
genome and other proteins and mRNAs. Pointing the computer
mouse at any feature allows the user to open a small menu that
links to additional descriptions.
70
Consider the new gene indicated by the red arrow in
Fig. 13.3. To view general information about this gene, hold the
computer mouse over the gene graphic and select Transcript
Information from the pop-up menu. The GeneView window
(Fig. 13.4) provides a description of this gene, as well as a link
to the GeneView window for the putative human ortholog
(Fig. 13.4, Homology Matches section). To view the database
sequences that align with the predicted exons of the new mouse
gene, place the computer mouse pointer over the gene in the
Detailed View (Fig. 13.3, arrow) and select Supporting evidence
from the pop-up menu. Fig. 13.5 depicts the mRNA and pro-
tein sequences that align with exons in the new gene. Click on
any of the green boxes to see the alignment of the database
sequence with the new transcript.
The zoomed-out Detailed View also provides links to com-
puted regions of orthology between the mouse and human
genomes (Fig. 13.3, pink bars). As the mouse genome assembly
and annotation lag behind those of the human, it may also
be useful to view the human genes in an orthologous region of
the genome.
UCSC also provides a mouse genome browser and the BLAT
search tool for use with the latest mouse genome sequence
assemblies. The links are available from the UCSC genome
browser home page, at http://genome.ucsc.edu/. Mouse
genome analysis tools developed at the NCBI, including a
mouse Map Viewer and mouse BLAST pages, are available
from http://www.ncbi.nlm.nih.gov/genome/guide/mouse/.
supplement to nature genetics • september 2002
 user’s guide
Figure 13.1
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 13.2

supplement to nature genetics • september 2002 71

 user’s guide
Figure 13.3
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

Figure 13.4

72 supplement to nature genetics • september 2002

 user’s guide
Figure 13.5
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

supplement to nature genetics • september 2002 73

 user’s guide
Commentary: keeping biology in mind
doi:10.1038/ng979
In working through the examples in the User’s Guide, the reader
is exposed to a number of databases, web sites and other
resources of enormous value for performing in silico analysis of
biological data. Familiarity with and use of this vast arsenal can
help the researcher to plan and execute experiments more intelli-
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics
gently. In using these resources and, more importantly, in draw-
ing biological conclusions based on the results gleaned from
these sites, however, there are a number of caveats and potential
pitfalls of which the user should be aware. Although some of the
specific points we now discuss go beyond the sample questions
included in this guide, the basic lessons to be learned apply to the
full range of bioinformatic analyses.
The user must understand the capabilities—and limitations—
of the programs being used. In the same way that molecular biol-
ogists need to understand the chemistry underlying a routine
assay or the physics behind separation techniques, they must
have a basic understanding of what search or analysis methods
actually do once the ‘Submit’ button has been pressed. Under-
standing what the chemistry, physics or search methods can and
cannot reveal is critical if the user is to extract the full meaning of
the results but not overinterpret them. By understanding the
methods, users can also optimize them and end up with a better
set of results than if these sequence-based search methods were
treated simply as a ‘black box’.
A specific case in which the reader could have encountered dif-
ficulty deals with the detection of domains within a protein, as
described in Question 10. Consider the part of the question that
discussed the Conserved Domain Database (CDD) at the NCBI.
The CDD is a ‘secondary database’, one in which the entries have
been derived from other databases, in this case Pfam23 and the
Simple Modular Architecture Research Tool (SMART)24. Pfam
provides collections of multiple sequence alignments that repre-
sent known, common protein domains. Pfam is subdivided into
two parts: Pfam A, which is manually curated, and Pfam B, which
is automatically generated. By virtue of being ‘hand-crafted’, the
entries in Pfam A are of higher quality and are therefore more
reliable than those in Pfam B. Nevertheless, both Pfam A and
Pfam B provide broad coverage across the spectrum of known
protein domains.
The second source database, SMART, provides information
on 500 domain families, but with a specific emphasis on those
domains that have been implicated in signaling or have been
found in extracellular or chromatin-associated proteins. This
was a deliberate choice by the developers, who wished to tackle
what might be called ‘tougher-to-detect’ or ‘tougher-to-define’
domains. At the outset, simply knowing the scope of the target
database tells the user whether or not it is an appropriate
choice for a sequence of interest, especially when some bio-
chemical data may already be available. If users were to search
solely against SMART and find nothing, without understand-
ing the limited scope of the data underlying the resource, they
might erroneously conclude that the protein of interest had no
known domains.
Continuing with this example, and assuming that the user
now understands the scope of the underlying source databases,
a second problem quickly surfaces. When searching Pfam and
SMART through the CDD interface at the NCBI, the search
is performed using a variation of the BLAST algorithm called
RPS-BLAST25. If one were, however, to go directly to the Pfam or
74
SMART web sites and issue the query there, the searches would
be performed using a very different algorithm, a hidden Markov
model26. Although a description of the two different methods is
beyond the scope of this discussion, it is important to understand
that they are fundamentally different and will therefore produce
different results. An extended discussion on this point, using spe-
cific examples, is available27. The CDD front end will miss those
SMART and Pfam entries that represent short domains, repeats
and motifs28. To understand what the methods do does not mean
having to comprehend advanced mathematical equations: basic
explanations in layman’s terms can be found in any one of a
number of reviews or textbooks7,8.
One can often carry out a search and become excited on the
identification of a motif; frequently such a motif is rather small.
The Lys-Asp-Glu-Leu motif is an example; it targets proteins to
the endoplasmic reticulum. But one should beware the ‘short-
motif ’ pitfall. The level of sequence identity required for signifi-
cant homology is much higher for smaller regions—they either
match or they don’t. For very short motifs, homology cannot be
inferred by sequence identity, meaning that short motifs may not
be at all helpful in describing what a protein does.
Longer motifs have greater power in identifying true positives
and eliminating false positives. More importantly, the support-
ing information is made available by simply clicking past the first
page of summary results provided by the search engine. Even, or
especially, the newest of users is encouraged to click away and
discover the information and assumptions underlying the results
that the searches have produced. These are self-explanatory in
many cases.
With respect to complete sequences, the reader is advised to
recall that the preliminary analyses of the human genome
sequence led to a large reduction in the estimated number of
genes contained in the human genome. Earlier, numbers of the
order of 80,000 to as high as 140,000 had been suggested29. With
the draft sequence of the genome in hand, new estimates lie
closer to 30,000–35,000 genes11. If this is correct, the human
would have only twice as many genes as are observed in either the
roundworm or the fruit fly11. At the same time, human genes
appear (in general) to have a more complex structure.
This pronounced ‘reduction’ in the number of genes in the
human genome obviously challenges the one-gene, one-protein
hypothesis (or, more properly, the one-gene, one-enzyme hypoth-
esis30), as the number of proteins in the human proteome is
thought to be well in excess of 35,000 (ref. 11). One explanation of
the large number of individual proteins that can be generated
from this relatively small number of genes is alternative splicing, a
process by which the transcripts from a single gene can be
processed differently and thus give rise to several distinct proteins.
Particularly germane to this discussion is that many proteins have
more than one function, depending on where they are found in
the cell or within the body as a whole.
An interesting example of this phenomenon is the multifunc-
tional protein phosphoglucose isomerase31. This protein cat-
alyzes the interconversion of D-glucose-6-phosphate and
D-fructose-6-phosphate. It is identical to neuroleukin, a protein
secreted by T cells that promotes the survival of some embryonic
spinal neurons and sensory nerves. It is also identical to an
autocrine motility factor that might be involved in metastasis,
and to a differentiation and maturation mediator implicated in
supplement to nature genetics • september 2002

the in vitro differentiation of human myeloid leukemia HL-60
cells to terminal monocytes. This therefore appears to be a single
soluble protein that can take on four distinct cellular roles.
A more extreme example of one protein being used in alterna-
tive contexts involves an outright phase shift: the proteins known
as α-enolase and τ-crystallin are encoded by a single gene and
have the same amino-acid sequence. In the liver, the protein
functions as α-enolase, a soluble glycolytic enzyme, whereas
within the lens of the eye, it functions as τ-crystallin, a structural
protein32. Proteins for which alternative functions have been
identified have been given the playful name ‘moonlighting pro-
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics
teins’ (see ref. 33 for a review).
Why is this biological finding important to anyone who uses
comparative sequence information? In the early days of sequence
comparison, it was assumed that if a sequence of unknown func-
tion matched a sequence of known function, one knew, by exten-
sion, the function of the unknown; the conclusions of many
published papers were based on this assumption. In light of these
and similar, more recent findings, does sequence similarity still
imply common function? The answer is: maybe yes and maybe
no. In any case, more evidence than just sequence similarity is
needed to draw any conclusion about sequence function.
Moving up in conceptual complexity to the level of structure,
an entire class of molecular modeling techniques is available to
consider similarities between proteins whose relationship might
not be obvious from looking strictly at the nucleotide or amino-
acid sequence. The reason one would want to perform such
analyses was stated early in a relatively short history of bioinfor-
matics34: structure is conserved to a greater extent than sequence.
This stands to reason, as there is evolutionary pressure to main-
tain the three-dimensional shape of proteins, particularly those
critical to the basic functions of a cell.
Inferring common function from structural similarity, how-
ever, is more problematic. Consider the TIM barrel. It defines a
structural superfamily whose members show a high degree of
structural similarity over a substantial number of residues. The
TIM-barrel fold is a good example of possible divergent evolu-
tion, because this same basic structure mediates a wide variety of
chemical reactions critical to biological survival. The TIM barrel
is associated with one non-enzymatic and fifteen enzymatic
functions35, and transcripts encoding TIM-barrel proteins
account for over 8% of the yeast transcriptome36. The roles of
TIM-barrel proteins are diverse, ranging from isomerases to
oxidoreductases and hydrolases. This generic versatility is eco-
nomical for the cell but can make the job of assigning function to
structures or substructures difficult. In deciding whether struc-
tural similarity implies common function, one needs to consider
the subcellular localization of the proteins, when they are
expressed, and the presence or absence of cofactors that might
significantly alter their structure.
A final point to be considered relates to annotations in the
public databases. Although these are of great value, most are
supplement to nature genetics • september 2002
user’s guide
made in an automated fashion, without the benefit of human
curation. This is a matter of practicality, as it would be difficult to
verify every annotation in the human genome, let alone those of
every sequenced organism. Although some sequence-based
annotations, such as the positions of genome, are determined
experimentally and are therefore quite reliable, others are no
more than predictions. The most notable of these are the predic-
tions of gene structure that can be found at the NCBI, Ensembl
and UCSC. Question 7 in this guide provides an excellent exam-
ple of inconsistencies in gene predictions obtained using meth-
ods; the user should use such information carefully, particularly
when designing experiments.
The second type of annotation—functional annotation—can
be even more problematic. Even when similarity can be reliably
detected, the functional annotations currently found in the pub-
lic databases are often incorrect. For example37, the functional
annotations of 340 Mycoplasma genes were assessed: 8% were
found to be incorrect, and, in many cases, did not logically con-
nect to the known biology and metabolism of Mycoplasma. So
never use database annotation as evidence of function when
there are few homologs or when the annotations are inconsistent
between homologs. And remember that annotations are intran-
sitive38: if protein A and protein B share a common functional
annotation, and so do proteins B and C, proteins A and C do not
necessarily have the same function. Use functional annotations
as a first step, and confirm the annotations by going back into the
primary literature.
Biology is complex, and we still do not understand it very well.
Although performing searches and finding data are not difficult,
the intelligent use of all of the accumulated facts from databases
is. It is always necessary to take a step backwards and ask a very
simple question: do the search results actually make biological
sense? Even when one is able to make biological sense of a predic-
tion of function, it may turn out to be incorrect. As science is
increasingly undertaken in a ‘sequence-based’ fashion, using
sequence data to underpin the experimental design and interpre-
tation of experiments, it becomes increasingly important that
computational results are cross-checked in the laboratory, against
the literature and with more robust computational analysis, so
that the conclusions not only make sense, but are also correct.
Acknowledgments
David Haussler (University of California, Santa Cruz), Ewan Birney
(The Wellcome Trust Sanger Institute) and David J. Lipman (National
Center for Biotechnology Information) served as advisors during the
development of this guide.
The authors would also like to thank the following people for their
contributions: K.N. Lazarides, S.K. Loftus, E.H. Margulies, K.L. Mohlke,
P.M. Pollock, R.B. Sood and J.W. Touchman (National Human Genome
Research Institute); D. Karolchik and J. Kent (University of California,
Santa Cruz); D. Church and K. Pruitt (National Center for Biotechnology
Information); and M. Hammond and E. Schmidt (Ensembl).
75

the in vitro differentiation of human myeloid leukemia HL-60
cells to terminal monocytes. This therefore appears to be a single
soluble protein that can take on four distinct cellular roles.
A more extreme example of one protein being used in alterna-
tive contexts involves an outright phase shift: the proteins known
as α-enolase and τ-crystallin are encoded by a single gene and
have the same amino-acid sequence. In the liver, the protein
functions as α-enolase, a soluble glycolytic enzyme, whereas
within the lens of the eye, it functions as τ-crystallin, a structural
protein32. Proteins for which alternative functions have been
identified have been given the playful name ‘moonlighting pro-
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics
teins’ (see ref. 33 for a review).
Why is this biological finding important to anyone who uses
comparative sequence information? In the early days of sequence
comparison, it was assumed that if a sequence of unknown func-
tion matched a sequence of known function, one knew, by exten-
sion, the function of the unknown; the conclusions of many
published papers were based on this assumption. In light of these
and similar, more recent findings, does sequence similarity still
imply common function? The answer is: maybe yes and maybe
no. In any case, more evidence than just sequence similarity is
needed to draw any conclusion about sequence function.
Moving up in conceptual complexity to the level of structure,
an entire class of molecular modeling techniques is available to
consider similarities between proteins whose relationship might
not be obvious from looking strictly at the nucleotide or amino-
acid sequence. The reason one would want to perform such
analyses was stated early in a relatively short history of bioinfor-
matics34: structure is conserved to a greater extent than sequence.
This stands to reason, as there is evolutionary pressure to main-
tain the three-dimensional shape of proteins, particularly those
critical to the basic functions of a cell.
Inferring common function from structural similarity, how-
ever, is more problematic. Consider the TIM barrel. It defines a
structural superfamily whose members show a high degree of
structural similarity over a substantial number of residues. The
TIM-barrel fold is a good example of possible divergent evolu-
tion, because this same basic structure mediates a wide variety of
chemical reactions critical to biological survival. The TIM barrel
is associated with one non-enzymatic and fifteen enzymatic
functions35, and transcripts encoding TIM-barrel proteins
account for over 8% of the yeast transcriptome36. The roles of
TIM-barrel proteins are diverse, ranging from isomerases to
oxidoreductases and hydrolases. This generic versatility is eco-
nomical for the cell but can make the job of assigning function to
structures or substructures difficult. In deciding whether struc-
tural similarity implies common function, one needs to consider
the subcellular localization of the proteins, when they are
expressed, and the presence or absence of cofactors that might
significantly alter their structure.
A final point to be considered relates to annotations in the
public databases. Although these are of great value, most are
supplement to nature genetics • september 2002
user’s guide
made in an automated fashion, without the benefit of human
curation. This is a matter of practicality, as it would be difficult to
verify every annotation in the human genome, let alone those of
every sequenced organism. Although some sequence-based
annotations, such as the positions of genome, are determined
experimentally and are therefore quite reliable, others are no
more than predictions. The most notable of these are the predic-
tions of gene structure that can be found at the NCBI, Ensembl
and UCSC. Question 7 in this guide provides an excellent exam-
ple of inconsistencies in gene predictions obtained using meth-
ods; the user should use such information carefully, particularly
when designing experiments.
The second type of annotation—functional annotation—can
be even more problematic. Even when similarity can be reliably
detected, the functional annotations currently found in the pub-
lic databases are often incorrect. For example37, the functional
annotations of 340 Mycoplasma genes were assessed: 8% were
found to be incorrect, and, in many cases, did not logically con-
nect to the known biology and metabolism of Mycoplasma. So
never use database annotation as evidence of function when
there are few homologs or when the annotations are inconsistent
between homologs. And remember that annotations are intran-
sitive38: if protein A and protein B share a common functional
annotation, and so do proteins B and C, proteins A and C do not
necessarily have the same function. Use functional annotations
as a first step, and confirm the annotations by going back into the
primary literature.
Biology is complex, and we still do not understand it very well.
Although performing searches and finding data are not difficult,
the intelligent use of all of the accumulated facts from databases
is. It is always necessary to take a step backwards and ask a very
simple question: do the search results actually make biological
sense? Even when one is able to make biological sense of a predic-
tion of function, it may turn out to be incorrect. As science is
increasingly undertaken in a ‘sequence-based’ fashion, using
sequence data to underpin the experimental design and interpre-
tation of experiments, it becomes increasingly important that
computational results are cross-checked in the laboratory, against
the literature and with more robust computational analysis, so
that the conclusions not only make sense, but are also correct.
Acknowledgments
David Haussler (University of California, Santa Cruz), Ewan Birney
(The Wellcome Trust Sanger Institute) and David J. Lipman (National
Center for Biotechnology Information) served as advisors during the
development of this guide.
The authors would also like to thank the following people for their
contributions: K.N. Lazarides, S.K. Loftus, E.H. Margulies, K.L. Mohlke,
P.M. Pollock, R.B. Sood and J.W. Touchman (National Human Genome
Research Institute); D. Karolchik and J. Kent (University of California,
Santa Cruz); D. Church and K. Pruitt (National Center for Biotechnology
Information); and M. Hammond and E. Schmidt (Ensembl).
75

Web resources: Internet resources featured in this guide
doi:10.1038/ng981
Major Genome Browsers
Ensembl
http://www.ensembl.org
NCBI Map Viewer
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics
http://www.ncbi.nlm.nih.gov/cgi-bin/Entrez/map_search
UCSC Genome Browser
http://genome.ucsc.edu
Additional Genome Browsers
In addition to the genome browsers discussed in this Guide, the
reader may find these additional views of the human genome
sequence helpful. Each of these sites provides documentation on
their scope of coverage and how to examine the data housed at
that site.
Celera
http://www.celera.com/genomics/academic/home.cfm
ORNL Genome Channel
http://compbio.ornl.gov/channel/
RIKEN Genomic Sciences Center
http://hgrep.ims.u-tokyo.ac.jp/
Genome annotation
The following sites provide detailed information on annotations The ability to bring together expressed sequence tag, mRNA and
at each of the three major genome portals.
Distributed Annotation System
http://www.ensembl.org/Docs/wiki/html/EnsemblDocs/
EnsemblDAS.html
Ensembl Science Documentation
http://www.ensembl.org/Docs/wiki/html/EnsemblDocs/
ScienceDocumentation.html
NCBI Contig Assembly and Annotation Process
http://www.ncbi.nlm.nih.gov/genome/guide/build.html
UCSC Annotation Database
http://genome.ucsc.edu/goldenPath/help/hg
TracksHelp.html
Human Genome Hub and Genome Central
These sites provide jumping-off points to major genome-based
web sites. Resources available include trace data archives, access
to cDNA and expressed sequence tag data and mapping informa-
tion used to produce genome assemblies. The web sites of the
individual members of the International Human Genome
Sequencing Consortium may be accessed through these sites.
Ensembl Human Genome Central
http://www.ensembl.org/genome/central/
NCBI Human Genome Central
http://www.ncbi.nlm.nih.gov/genome/guide/central.html
supplement to nature genetics • september 2002
user’s guide
NHGRI Genome Hub
http://www.nhgri.nih.gov/genome_hub.html
UK HGMP GenomeWeb
http://www.hgmp.mrc.ac.uk/GenomeWeb/genome-
db.html
Major public sequence databases
Each of these databases belongs to the International Nucleotide
Sequence Database Collaboration. Although all three centers
provide separate mechanisms for sequence submission by indi-
vidual investigators, they exchange data daily. As each member
database stores and presents the underlying data using a slightly
different format, this data exchange makes all known nucleotide
and protein sequence data available to all users, regardless of
which of the three databases are queried.
DNA Data Bank of Japan
http://www.ddbj.nig.ac.jp
EMBL Nucleotide Sequence Database
http://www.ebi.ac.uk/embl/index.html
GenBank
http://www.ncbi.nlm.nih.gov
Expressed sequence tag clustering databases
other related sequences into gene-oriented clusters often facili-
tates genomic analysis, since the method groups individual
sequences that most likely arise from the same gene or tran-
script. These three databases provide gene-oriented views of the
data, using different algorithms in calculating the individual
gene clusters.
STACK
http://www.sanbi.ac.za/Dbases.html
TIGR Gene Indices
http://www.tigr.org/tdb/tgi.shtml
UniGene
http://www.ncbi.nlm.nih.gov/UniGene
Human genetic and physical maps
The databases listed below represent a significant portion of the
data underlying current human genome assemblies. Many of
these data are available through DDBJ/EMBL/GenBank, but each
database contains additional information regarding clones, con-
structs and similar that is not available through the major
sequence repositories. A more extensive list of human genetic
and physical maps can also be found through the online Nucleic
Acids Research Database Collection, at http://nar.oupjournals.
org/cgi/content/ full/30/1/1/DC1.
Bacterial artificial chromosome and accession maps
http://genome.wustl.edu/projects/human/index.php?fpc=1
77
 user’s guide
GenAtlas Ensembl BLAST
http://www.citi2.fr/GENATLAS/ http://www.ensembl.org/Homo_sapiens/blastview

Genebridge4 radiation hybrid maps SSAHA

http://www.sanger.ac.uk/Software/RHserver/ http://www.ensembl.org/Homo_sapiens/ssahaview
RHserver.shtml
Model organism databases
GeneMap ’99 This list represents a small subset of the sequencing initiatives on
http://www.ncbi.nlm.nih.gov/genemap99 model organisms. Additional information on the progress of
numerous model organism sequencing initiatives can be found
GenMapDB on the Model Organisms for Biomedical Research web page, at
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

http://genomics.med.upenn.edu/genmapdb http://www.nih.gov/science/models/. A more extensive list of

organismal databases can also be found through the online
Généthon linkage map Nucleic Acids Research Database Collection, at http://nar.
http://www.genethon.fr/index_en.html oupjournals.org/cgi/content/full/30/1/1/DC1.

HuGeMap Arabidopsis thaliana

http://www.infobiogen.fr/services/Hugemap The Arabidopsis Information Resource
http://www.arabidopsis.org
Marshfield genetic maps
http://research.marshfieldclinic.org/genetics/ Arabidopsis Genome Initiative
Map_Markers/maps/IndexMapFrames.html http://mips.gsf.de/proj/thal/db/

RHdb Caenorhabditis elegans

http://corba.ebi.ac.uk/RHdb AceDB
http://www.acedb.org
Stanford G3 and TNG radiation hybrid maps
http://www-shgc.stanford.edu/RH/ WormBase
http://www.wormbase.org/
Genomic Databases and Resources
In addition to the databases listed in the section above, there are Drosophila melanogaster
numerous useful databases containing human mutation, varia- Berkeley Drosophila Genome Project
tion, medical or expression data. This short list is offered as a rep- http://www.fruitfly.org/
resentative cross-section of the types of database freely available
to genome researchers. The reader is referred to the ‘lists of lists’ FlyBase
found at the Human GenomeHub and Genome Central cites for http://flybase.bio.indiana.edu/
a more extensive catalog of available resources.
Escherichia coli
Cancer Genome Anatomy Project (CGAP) EcoGene
http://www.ncbi.nlm.nih.gov/CGAP/ http://bmb.med.miami.edu/EcoGene/EcoWeb/

Genome DataBase (GDB) Microbial Genomes

http://www.gdb.org Comprehensive Microbial Resource
http://www.tigr.org/tigr-scripts/CMR2/
HUGO Gene Nomenclature CMRHomePage.spl
http://www.gene.ucl.ac.uk/nomenclature
TIGR Microbial Database
Online Mendelian Inheritance in Man (OMIM) http://www.tigr.org/tdb/mdb/
http://www.ncbi.nlm.nih.gov/Omim
Mouse
SNP Consortium Mouse Genome Database/Informatics
http://snp.cshl.org http://www.informatics.jax.org/

Sequence-based searching Rat

The following links provide access to the most frequently used Rat Genome Database
tools for performing sequence-based comparisons to human http://rgd.mcw.edu
genome data. An extensive list of sequence similarity search
tools can be found on the ExPASy web site, at http://us.expasy. Yeast
org/tools/. Comprehensive Yeast Genome Database
http://mips.gsf.de/proj/yeast/CYGD/db/
BLAST
http://www.ncbi.nlm.nih.gov/BLAST/ Saccharomyces Genome Database
http://genome-www.stanford.edu/Saccharomyces/
BLAT
http://genome.ucsc.edu/cgi-bin/hgBlat?command=start

78 supplement to nature genetics • september 2002


S. pombe Genome Sequencing Project
http://www.sanger.ac.uk/Projects/S_pombe/
Zebrafish
Zebrafish Information Network
http://zfin.org
user’s guide
Genetic education
The following sites present basic information on genetics and
genomics, much of which is appropriate for elementary and sec-
ondary school education, as well as for the college level. Many of
these sites offer teaching plans, graphics and other teaching
resources that can be freely used in the classroom or lecture hall.

Ethical, legal and social Issues Access Excellence

Although this guide has focused on the mechanics of accessing http://www.accessexcellence.org/
and using human genome data, it is important to remember that
ethical, legal and social issues (ELSI) are becoming increasingly Department of Energy education resources
© 2002 Nature Publishing Group http://www.nature.com/naturegenetics

important in this age of genetic and genomic research. The fol- http://www.ornl.gov/hgmis/education/education.html
lowing web sites provide an introduction to important issues
related to genome biology as applied to human health and pro- Genetics Education Center
vide a jumping-off point for further information. http://www.kumc.edu/gec/

DOE ELSI Program NHGRI Exploring our Molecular Selves Multimedia Kit
http://www.ornl.gov/hgmis/elsi/elsi.html http://www.genome.gov/Pages/EducationKit/

Lawrence Berkeley National Laboratory NHGRI Glossary of Genetic Terms

http://www.lbl.gov/Education/ELSI/ http://www.genome.gov/glossary.cfm

NHGRI ELSI Program

http://www.nhgri.nih.gov/ELSI/

supplement to nature genetics • september 2002 79