Abstract

Advances in many areas of genomics research are heavily rooted in engineering technology, from the capillary electrophoresis units used in large-scale DNA sequencing projects, to the photolithography and robotics technology used in chip manufacture, to the confocal imaging systems used to read those chips, to the beam and detector technology driving high-throughput mass spectroscopy. Further advances in (for example) materials science and nanotechnology promise to improve the sensitivity and cost of these technologies greatly in the near future. Genomic research makes it possible to look at biological phenomena on a scale not previously possible: all genes in a genome, all transcripts in a cell, all metabolic processes in a tissue. One feature that all of these approaches share is the production of massive quantities of data. GenBank, for example, now accommodates >10 10 nucleotides of nucleic acid sequence data and continues to more than double in size every year. New technologies for assaying gene expression patterns, protein structure, protein-protein interactions, etc., will provide even more data. How to handle these data, make sense of them, and render them accessible to biologists working on a wide variety of problems is the challenge facing bioinformatics--an emerging field that seeks to integrate computer science with applications derived from molecular biology. We are swimming in a rapidly rising sea of data...how do we keep from drowning? Bioinformatics faces its share of growing pains, many of which presage problems that all biologists will soon encounter as we focus on large-scale science projects. For starters, few scientists can claim a strong background on both sides of the divide separating computer science from biomedical research. This shortage means a lack of mentors who might train the next generation of "bioinformaticians." Lack of familiarity with the intellectual questions that motivate each side can also lead to misunderstandings. For example, writing a computer program that assembles overlapping expressed sequence tag (EST) sequences may be of great importance to the biologist without breaking any new ground in computer science. Similarly, proving that it is impossible to determine a globally optimal phylogenetic tree under certain conditions may constitute a significant finding in computer science, while being of little practical use to the biologist. Identifying problems of intellectual value to all concerned is an important goal for the maturation of computational biology as a distinct discipline. "Real" biology is increasingly carried out in front of a computer, while an increasing number of projects in computer science will be driven by biological problems. Further difficulties stem from the fact that bioinformatics is an inherently integrative discipline, requiring access to data from a wide range of sources. Without the underlying data, and the ability to combine these data in new and interesting ways, the field of bioinformatics would be very much limited in scope. For example, the widespread utility of BLAST for the identification of gene similarity (1) is attributable not only to the algorithm itself (and its implementation), but also to the availability of databases such as GenBank, the European Molecular Biology Laboratory (EMBL), and the DNA Data Bank of Japan (DDBJ), which pool genomic data from a variety of sources. BLAST would be of limited utility without a broad-based database to query. One core aspect of research in computational biology focuses on database development: how to integrate and optimally query data from (for example) genomic DNA sequence, spatial and temporal patterns of mRNA expression, protein structure, immunological reactivity, clinical outcomes, publication records, and other sources. A second focus involves pattern recognition algorithms for such areas as nucleic acid or protein sequence assembly, sequence alignment for similarity comparisons or phylogeny reconstruction, motif recognition in linear sequences or higher-order structure, and common patterns of gene expression. Both database integration and pattern recognition depend absolutely on accessing data from diverse sources, and being able to integrate, transform, and reproduce these data in new formats. As noted above, computational biology is a fundamentally collaborative discipline, owing its very existence to the availability of rich and extensive data sets for analysis, integration, and manipulation. Data accessibility and usability are therefore critical, raising concerns about data release policies--what constitutes primary data, who owns this resource, when and how data should be released, and what restrictions may be placed on further use. Two challenges have emerged that could potentially restrict the advancement of bioinformatics research: (i) questions related to the appropriate use of data released before publication and (ii) restrictions on the reposting of published data. The first challenge to bioinformatics research relates to the analysis of data posted on the Web in advance of publication. Recognizing the value of early data release for a wide range of studies, the Human Genome Project adopted a policy of prepublication data release (2), and many genome projects (and the funding agencies that support them) now adhere to similar rules. Because bioinformatics depends absolutely on the ability to integrate data from a wide variety of sources, it is to be hoped that other projects that generate genomic-scale data (including expression analysis and proteomics research) will follow a similar policy (3), because immensely valuable results can emerge from large-scale comparative studies of genome structure, microarray data, protein interactions, and so on (4-6). The success of such altruistic data release policies, however, requires that those who generate primary sequence data (often on behalf of the community at large) receive appropriate recognition and are able to derive intellectual satisfaction from their work. Rowen et al. (7) have recently proposed treating unpublished data available on the Web as analogous to "personal communication," thereby establishing some degree of intellectual property protection.