Clinical review

Science, medicine, and the future

Bioinformatics
Ardeshir Bayat

Bioinformatics, An unprecedented wealth of biological data has been

a new generated by the human genome project and sequenc- Summary points
interdisciplinary ing projects in other organisms. The huge demand for
science, is analysis and interpretation of these data is being man-
Bioinformatics is the application of tools of
essential to aged by the evolving science of bioinformatics.
computation and analysis to the capture and
managing, Bioinformatics is defined as the application of tools of
interpretation of biological data
understanding, computation and analysis to the capture and interpret-
and harnessing ation of biological data. It is an interdisciplinary field, Bioinformatics is essential for management of
clinical benefit which harnesses computer science, mathematics, phys- data in modern biology and medicine
from new ics, and biology (fig 1). Bioinformatics is essential for
genetic data management of data in modern biology and medicine. The bioinformatics toolbox includes computer
This paper describes the main tools of the bioinforma- software programs such as BLAST and Ensembl,
Centre for
tician and discusses how they are being used to which depend on the availability of the internet
Integrated Genomic interpret biological data and to further understanding
Medical Research, of disease. The potential clinical applications of these
University of
Analysis of genome sequence data, particularly
Manchester, data in drug discovery and development are also the analysis of the human genome project, is one
Manchester discussed. of the main achievements of bioinformatics to
M13 9PT
date
Ardeshir Bayat
MRC fellow Methods
Prospects in the field of bioinformatics include its
Correspondence to: This article is based on personal experience in
ardeshir.bayat@ future contribution to functional understanding
man.ac.uk bioinformatics and on selected articles in recent issues of the human genome, leading to enhanced
of Nature Genetics, Nature Genetics Reviews, Nature Medi- discovery of drug targets and individualised
BMJ 2002;324:1018–22 cine, and Science. Key terms including bioinformatics, therapy
comparative and functional genomics, proteomics,
microarray, disease, and medicine were used to search
for relevant articles in the peer reviewed scientific
literature. the worldwide human genome project and a private
genomic company.1 2 However, in recent years, the sci-
Bioinformatics and its impact on entific world has witnessed the completion of whole
genomics genome sequences of many other organisms. The
analysis of the emerging genomic sequence data and
Last year it was announced that the entire human the human genome project is a landmark achievement
genome had been mapped as a result of the efforts of for bioinformatics.
A novel strategy for random sequencing of the
whole genome (the so called “shot gun” technique) was
used to sequence the genome of Haemophilus influenzae
Biology Medicine in 1995.3 This was the very first complete genome of
any free living organism to be sequenced. Other bacte-
rial genomes, such as those of Mycoplasma genitalium
Bioinformatics and Mycobacterium tuberculosis, were sequenced soon
after,4 5 and the sequence of the plague bacterium Yers-
inia pestis was recently completed.6 The sequence and
Maths/physics Computer science annotation of the first eukaryotic genome, that of Sac-
charomyces cerevisiae (a yeast),7 was followed by those of
other eukaryotic species such as Caenorhabtidis elegans
An additional figure
(a worm),8 Drosophila melanogaster (fruit fly),9 and Arab-
appears on Fig 1 Interaction of disciplines that have contributed to the dopsis thaliana (mustard weed)10 (see fig A on bmj.com).
bmj.com formation of bioinformatics Sequencing of several other species, including

1018 BMJ VOLUME 324 27 APRIL 2002 bmj.com

 Clinical review

researcher, beyond a basic acquisition and analysis of

Useful bioinformatic websites (available freely simple data, would certainly need external bioinfor-
on the internet) matic advice for any complex analysis.
• National Center for Biotechnology Information The growth of bioinformatics has been a global
(www.ncbi.nlm.nih.gov)—maintains bioinformatic tools venture, creating computer networks that have allowed
and databases easy access to biological data and enabled the develop-
• National Center for Genome Resources ment of software programs for effortless analysis.
(www.ncgr.org/)—links scientists to bioinformatics Multiple international projects aimed at providing
solutions by collaborations, data, and software gene and protein databases are available freely to the
development whole scientific community via the internet.
• Genbank (www.ncbi.nlm.nih.gov/Genbank)—stores
and archives DNA sequences from both large scale
genome projects and individual laboratories Bioinformatic analysis
• Unigene (www.ncbi.nlm.nih.gov/UniGene)—gene
sequence collection containing data on map location
The escalating amount of data from the genome
of genes in chromosomes projects has necessitated computer databases that fea-
• European Bioinformatic Institute ture rapid assimilation, usable formats and algorithm
(www.ebi.ac.uk)—centre for research and services in software programs for efficient management of
bioinformatics; manages databases of biological data biological data.13 Because of the diverse nature of
• Ensembl (www.ensembl.org)—automatic annotation emerging data, no single comprehensive database
database on genomes exists for accessing all this information. However, a
• BioInform (www.bioinform.com)—global growing number of databases that contain helpful
bioinformatics news service information for clinicians and researchers are avail-
• SWISS-PROT (www.expasy.org/sprot/)—important able. Information provided by most of these databases
protein database with sequence data from all is free of charge to academics, although some sites
organisms, which has a high level of annotation
require subscription and industrial users pay a licence
(includes function, structure, and variations) and is
minimally redundant (few duplicate copies) fee for particular sites. Examples range from sites pro-
• International Society for Computational Biology
viding comprehensive descriptions of clinical disor-
(www.iscb.org/)—aims to advance scientific ders, listing disease susceptibility genetic mutations
understanding of living systems through computation; and polymorphisms, to those enabling a search for dis-
has useful bioinformatic links ease genes given a DNA sequence (box).
These databases include both “public” repositories
of gene data as well as those developed by private com-
zebrafish, pufferfish, mouse, rat, and non-human panies. The easiest way to identify databases is by
primates, are either under way or nearing completion
by both private and public sequencing initiatives.11 The
knowledge obtained from these sequence data will
have considerable implications for our understanding
of biology and medicine. As a result of comparative
genomic and proteomic research, we will soon be able
to not only locate each human gene but also fully
understand its function.12

Bioinformatic tools
The main tools of a bioinformatician are computer
software programs and the internet. A fundamental
activity is sequence analysis of DNA and proteins using
various programs and databases available on the world
wide web. Anyone, from clinicians to molecular
biologists, with access to the internet and relevant web-
sites can now freely discover the composition of
biological molecules such as nucleic acids and proteins
by using basic bioinformatic tools. This does not imply
that handling and analysis of raw genomic data can
easily be carried out by all. Bioinformatics is an evolv-
ing discipline, and expert bioinformaticians now use
complex software programs for retrieving, sorting out,
analysing, predicting, and storing DNA and protein
sequence data.
Large commercial enterprises such as pharmaceu-
tical companies employ bioinformaticians to perform
and maintain the large scale and complicated bioinfor-
Fig 2 Ensembl website: a genomic data search facility freely available on the internet.
matic needs of these industries. With an ever-
Ensembl is a joint project between the European Bioinformatic Institute and the Sanger
increasing need for constant input from bioinformatic Centre, which is capable of automatically tracking the sequenced pieces of the human
experts, most biomedical laboratories may soon have genome and assembling and analysing them to identify genes and other features of interest
their own in-house bioinformatician. The individual to biomedical researchers

BMJ VOLUME 324 27 APRIL 2002 bmj.com 1019

Clinical review

database links and searchable indexes provided by one

of the major public databases. For example, the
National Center for Biotechnology Information
(www.ncbi.nlm.nih.gov) provides the Entrez browser,
which is an integrated database retrieval system that
allows integration of DNA and protein sequence data-
bases. The European Bioinformatic Institute archives
gene and protein data from genome studies of all
organisms, whereas Ensembl produces and maintains
automatic annotation on eukaryotic genomes (fig 2).
The quality and reliability of databases vary; certainly
some of the better known and more established ones,
such as those above, are superior to others.
One of the simplest and better known search tools
is called BLAST (basic local alignment search tool, at
www.ncbi.nlm.nih.gov/BLAST/). This algorithm soft-
ware is capable of searching databases for genes with
similar nucleotide structure (fig 3) and allows compari-
son of an unknown DNA or amino acid sequence with
hundreds or thousands of sequences from human or
other organisms until a match is found. Databases of
known sequences are thus used to identify similar
sequences, which may be homologues of the query
sequence. Homology implies that sequences may be
related by divergence from a common ancestor or
share common functional aspects. When a database is
searched with a newly determined sequence (the query
Fig 3 Web page illustrating freely available BLAST services run by
the National Center for Biotechnology Information. BLAST (basic local
sequence), local alignment occurs between the query
alignment search tool) is a set of similarity search programs sequence and any similar sequence in the database.
designed to explore all of the available DNA sequence databases The result of the search is sorted in order of priority on
the basis of maximum similarity. The sequence with the
searching for bioinformatic tools and databases in any highest score in the database of known genes is the
one of the commonly used search engines. Another homologue. If homologues or related molecules exist
way to identify bioinformatic sources is through for a query sequence, then a newly discovered protein
may be modelled and the gene product may be
predicted without the need for further laboratory
Genomics
I experiments.
DNA ...agcttgatattatacgcgcggca
Functional genomics
C
Since the completion of the first draft of the human
o Makes
genome,1 2 the emphasis has been changing from
m
genes themselves to gene products. Functional genom-
p II ics assigns functional relevance to genomic infor-
l Transcriptomics mation. It is the study of genes, their resulting proteins,
e and the role played by the proteins.
x RNA Analysis and interpretation of biological data con-
i siders information not only at the level of the genome
t but at the level of the proteome and the transcriptome
y Makes (fig 4). Proteomics is the analysis of the total amount of
proteins (proteome) expressed by a cell, and transcrip-
III tomics refers to the analysis of the messenger RNA
IV transcripts produced by a cell (transcriptome). DNA
Proteomics
microarray technology determines the expression level
of genes and includes genotyping and DNA sequenc-
Protein
ing. Gene expression arrays allow simultaneous analy-
sis of the messenger RNA expression levels of
thousands of genes in benign and malignant tumours,
such as keloid and melanoma. Expression profiles clas-
Fig 4 Schematic diagram representing complexity of genomic data processing. Analysis and sify tumours and provide potential therapeutic
interpretation of biological data considers information at every level from the genome (total targets.14
genetic content) to the proteome (total protein content) and transcriptome (total messenger Bioinformatic protein research draws on annotated
RNA content) of the cell. The images numbered I-IV to the right of the diagram represent
protein and two dimensional electrophoresis data-
relevant examples of DNA (image I is base pair nucleotides); RNA (image II is a microarray
showing levels of gene expression); and protein (image III is a structure of a single protein; bases. After separation, identification, and characterisa-
image IV is a two dimensional gel electrophoresis showing separation of all proteins of a tion of a protein, the next challenge in bioinformatics is
cell—each spot corresponds to a different protein chain) the prediction of its structure. Structural biologists also

1020 BMJ VOLUME 324 27 APRIL 2002 bmj.com

 Clinical review

use bioinformatics to handle the vast and complex data

from x ray crystallography, nuclear magnetic reso- Additional educational resources
nance, and electron microscopy investigations to create
three dimensional models of molecules.15 Journals
• Specific bioinformatic journals exist (for example,
www.bioinformatics.oupjournals.org), but papers from
every area of science and medicine involving
Other applications of bioinformatics bioinformatic analysis are published in any biomedical
Apart from analysis of genome sequence data, journal. Examples include:
bioinformatics is now being used for a vast array of • The human genome (special issue). Nature
2001;409:813-933.
other important tasks, including analysis of gene varia-
tion and expression, analysis and prediction of gene • The human genome (special issue). Science
2001;5507:1145-434.
and protein structure and function, prediction and
• The human genome (special issue). JAMA
detection of gene regulation networks, simulation envi-
2001;286:2211-333.
ronments for whole cell modelling, complex modelling
• The human genome (special issue). Scientific
of gene regulatory dynamics and networks, and American 2000;283:38-57.
presentation and analysis of molecular pathways in • Luscombe NM, Greenbaum D, Gerstein M. What is
order to understand gene-disease interactions.16 bioinformatics? Method Inform Med 2001;40:346-58.
Although on a smaller scale, simpler bioinformatic • Online Lectures on Bioinformatics
tasks valuable to the clinical researcher can vary from (www.lectures.molgen.mpg.de/)
designing primers (short oligonucleotide sequences
Books
needed for DNA amplification in polymerase chain
• Mount DW. Bioinformatics: sequence and genome
reaction experiments) to predicting the function of analysis. Cold Spring Harbor Laboratory Press, 2001.
gene products.
• Baxevanis AD, Ouellette BFF. Bioinformatics: a
practical guide to the analysis of genes and proteins. 2nd ed.
John Wiley and Sons, 2001.
Clinical application of bioinformatics • Lengauer T (Ed). Bioinformatics. Wiley-VCH Series,
2001. (Methods and principles in medicinal chemistry
The clinical applications of bioinformatics can be
series.)
viewed in the immediate, short, and long term. The
• Higgins D, Taylor W. Bioinformatics. Oxford
human genome project plans to complete the human University Press, 2000. (Practical approach series.)
sequence by 2003, producing a database of all the vari- • Baldi P, Brunak S. Bioinformatics. 2nd ed. MIT Press,
ations in sequence that distinguish us all. The project 2001. (Adaptive computation and machine learning
could have considerable impact on people living in series.)
2020—for example, a complete list of human gene
BMJ archive
products may provide new drugs and gene therapy for
• Aitman TJ. DNA microarrays in medical practice.
single gene diseases may become routine BMJ 2001;323:611-5.
(www.ornl.gov/hgmis/medicine/tnty.html). • Mathew CG. Postgenomic technologies: hunting the
Basic bioinformatic tools are already accessed in genes for common disorders. BMJ 2001;322:1031-4.
certain clinical situations to aid in diagnosis and treat- • Stewart A, Haites N, Rose P. Online medical genetics
ment plans. For example, PubMed (www.nlm.nih.gov) resources: a UK perspective. BMJ 2001;322:1037-9.
is accessed freely for biomedical journals cited in • Savill J. Molecular genetic approaches to
Medline, and OMIM (Online Mendelian Inheritance in understanding disease. BMJ 1997;314:126-9.
Man at www3.ncbi.nlm.nih.gov/Omim/), a search tool
for human genes and genetic disorders, is used by cli-
nicians to obtain information on genetic disorders in In the longer term, integrative bioinformatic analy-
the clinic or hospital setting. An example of the appli- sis of genomic, pathological, and clinical data in clinical
cation of bioinformatics in new therapeutic advances is trials will reveal potential adverse drug reactions in
the development of novel designer targeted drugs such individuals by use of simple genetic tests. Ultimately,
as imatinib mesylate (Gleevec), which interferes with pharmacogenomics (using genetic information to
the abnormal protein made in chronic myeloid individualise drug treatment) is likely to bring about a
leukaemia.17 (Imatinib mesylate was synthesised at new age of personalised medicine; patients will carry
Novartis Pharmaceuticals by identifying a lead in a gene cards with their own unique genetic profile for
high throughput screen for tyrosine kinase inhibitors certain drugs aimed at individualised therapy and tar-
and optimising its activity for specific kinases.) The geted medicine free from side effects.
ability to identify and target specific genetic markers by
using bioinformatic tools facilitated the discovery of
this drug.
Future directions
In the short term, as a result of the emerging bioin- The practice of studying genetic disorders is changing
formatic analysis of the human genome project, more from investigation of single genes in isolation to
disease genes will be identified and new drug targets discovering cellular networks of genes, understanding
will be simultaneously discovered. Bioinformatics will their complex interactions, and identifying their role in
serve to identify susceptibility genes and illuminate the disease.19 As a result of this, a whole new age of
pathogenic pathways involved in illness, and will there- individually tailored medicine will emerge. Bioinfor-
fore provide an opportunity for development of matics will guide and help molecular biologists and
targeted therapy. Recently, potential targets in cancers clinical researchers to capitalise on the advantages
were identified from gene expression profiles.18 brought by computational biology.20 The clinical

BMJ VOLUME 324 27 APRIL 2002 bmj.com 1021

Clinical review

research teams that will be most successful in the com- 8 The C. elegans Sequencing Consortium. Genome sequence of the nema-
tode C. elegans: a platform for investigating biology. Science
ing decades will be those that can switch effortlessly 1998;282:2012-8.
between the laboratory bench, clinical practice, and the 9 Myers EW, Sutton GG, Delcher AL, Dew IM, Fasulo DP, Flanigan MJ, et al.
A whole-genome assembly of Drosophila. Science 2000;287:2196-204.
use of these sophisticated computational tools.
10 Arabidopsis Genomics Initiative. Analysis of the genome sequence of the
flowering plant Arabidopsis thaliana. Nature 2000;408:796-815.
I thank Tessa Richards, Dipak Roy, and Professor Bill Ollier for 11 Stein L. Genome annotation: from sequence to biology. Nat Rev Genet
advice on the preparation of this manuscript and Andy Brass for 2001;2:493-503.
providing me with some of the diagrams. 12 Subramanian G, Adams MD, Venter JC, Broder S. Implications of the
Funding: Medical Research Council. human genome for understanding human biology and medicine. JAMA
Competing interests: None declared. 2001;286:2296-306.
13 Benton D. Bioinformatics—principles and potential of a new multidisci-
plinary tool. Trends Biotech 1996;14:261-312.
1 International Human Genome Sequencing Consortium. Initial sequenc- 14 Maggio ET, Ramnarayan K. Recent developments in computational pro-
ing and analysis of the human genome. Nature 2001;409:860-921. teomics. Trends Biotech 2001;19:266-72.
2 Venter JC, Adams MD, Myers EW, Li PW, Mural RJ, Sutton GG, et al. The 15 Burley SK, Almo SC, Bonanno JB, Capel M, Chance MR, Gaasterland T,
sequence of the human genome. Science 2001;291:1304-51.
et al. Structural genomics: beyond the human genome project. Nat Genet
3 Fleischmann RD, Adams MD, White O, Clayton RA, Kirkness EF, Kerlav-
1999;23:151-7.
age AR, et al. Whole-genome random sequencing and assembly of Hae-
16 Tsoka S, Ouzounis CA. Recent developments and future directions in
mophilus influenzae Rd. Science 1995;269:496-512.
computational genomics. FEBS Lett 2000;480:42-8.
4 Fraser CM, Gocayne JD, White O, Adams MD, Clayton RA, Fleischmann
RD, et al. The minimal gene complement of Mycoplasma genitalium. 17 Druker BJ, Sawyers CL, Kantarjian H, Resta DJ, Reese SF, Ford JM, et al.
Science 1995;270:397-403. Activity of a specific inhibitor of the BCR-ABL tyrosine kinase in the blast
5 Cole ST, Brosch R, Parkhill J, Garnier T, Churcher C, Harris D, et al. Deci- crisis of chronic myeloid leukemia and acute lymphoblastic leukemia
phering the biology of Mycobacterium tuberculosis from the complete with the Philadelphia chromosome. N Engl J Med 2001;344:1038-42.
genome sequence. Nature 1998;393:537-44. 18 Graeber TG, Eisenberg D. Bioinformatic identification of potential auto-
6 Parkhill J, Wren BW, Thomson NR, Titball RW, Holden MT, Prentice MB, crine signaling loops in cancers from gene expression profiles. Nat Genet
et al. Genome sequence of Yersinia pestis, the causative agent of plague. 2001;29:295-300.
Nature 2001;413:523-27. 19 Debouk C, Metcalf B. The impact of genomics on drug discovery. Annu
7 Goffeau A, Barrell BG, Bussey H, Davis RW, Dujon B, Feldmann H, et al. Rev Pharmacol Toxicol 2000;40:193-208.
Life with 6000 genes. Science 1996;274:546. 20 Butler D. Are you ready for the revolution? Nature 2001;409:758-60.

When I use a word

Meta-

Mr John Gleave, a neurosurgeon, has written to ask me early 17th century (John Donne, for example, writes
the origin of the meta- in meta-analysis. The answer about metatheology) but did not become really
comes from Aristotle. popular until the middle of the 19th century. Examples
The Greek preposition ìåôá́ (meta) had several include metaethics (the study of the foundations of
meanings, depending on whether it governed the ethics, especially the nature of ethical statements) and
accusative, genitive, or dative case. With the accusative metahistory (an inquiry into the principles that govern
it could mean coming into or among, in pursuit of, or historical events).
coming after in place or time; with the genitive it could Then, from about 1940, it became commonplace to
mean in the midst of, between, or in common with; prefix meta- to designate concern with basic principles.
and with the dative it could mean in the company of or A metacriterion is a criterion that defines criteria. A
over and above. It was also used as a prefix to express metatheorem is a theorem about theorems. A
such notions as sharing, being in the midst of, metalanguage is a language that supplies terms for
succession, pursuit, reversal, and (most commonly) analysing a language; a metametalanguage does the
change. Examples of the last include metabolism, same for a metalanguage. And Jean Tinguely described
metamorphosis, and metaplasia. his machine-like sculptures as “metamechanical.” (But
In scientific English words its uses include a metaphysician is not a doctor’s doctor.)
“consequent upon” (as in the obsolete terms In these poststructuralist times we recognise many
meta-arthritic, metapneumonic), “behind” or “beyond” metaforms. Mantissa, a medical novel by John Fowles, is
in an anatomical sense (metabranchial, metacarpal, metafiction; Francois Truffaut’s film La Nuit Amercaine
metaphysis), “coming later” (metaphase, which comes is metacinema; several paintings by Magritte, notably
after prophase), or “changing” (metachromasia, a La Condition Humaine, are meta-art; and John Cage’s
property of materials that stain a different colour from piano piece 4’33’’ is metamusic.
the stain used). In geology meta- is used to distinguish So meta-analysis is an analysis of analyses, in which
various types of metamorphic processes. And chemists sets of previously published (or unpublished) data are
use meta- to differentiate certain metameric chemical themselves subjected as a whole to further analysis. In
compounds (such as metacresol, paracresol, this statistical sense it was first used in the 1970s by GV
orthocresol). Glass (Educ Res 1976;3(Nov):2). As he wrote, “The term
And so to Aristotle. Some 250 years after his death, is a bit grand, but it is precise and apt.” Incidentally,
Aristotle’s manuscripts came into the hands of meta-analysis should not be confused with metanalysis,
Andronicus of Rhodes, who edited them. Andronicus which is the process whereby, for example, “a nadder”
called one set of papers The Physics (ôá̀ *
ïõóéêá́), dealing becomes “an adder” (see BMJ 1999;318:1758 and
as they did with natural science. Then he published a 2000;321:953).
set of papers that he called The Metaphysics (ôá̀ ìåôá̀ ôá̀ I trust that this cures Mr Gleave’s metagrobolism.
*ïõóéêá́), simply because it came after The Physics.
Jeff Aronson clinical pharmacologist, Oxford
However, because The Metaphysics dealt with what
Aristotle called “primary philosophy,” or ontology, We welcome articles up to 600 words on topics such as
metaphysics came to be misunderstood as “the science A memorable patient, A paper that changed my practice, My
of that which transcends the physical.” most unfortunate mistake, or any other piece conveying
As a result, the prefix meta- was then used to instruction, pathos, or humour. If possible the article
designate any higher science (actual or hypothetical) should be supplied on a disk. Permission is needed
that dealt with more fundamental problems than the from the patient or a relative if an identifiable patient is
original science itself. This use first appeared in the referred to.

1022 BMJ VOLUME 324 27 APRIL 2002 bmj.com