Bioinformatics

Tao Jiang, Ying Xu, and Michael Q. Zhang, eds. Current Topics in Computational Molecular Biology. Cambridge: MIT Press, 2002. Pp. 542. $55.Thomas Lengauer, ed. Bioinformatics: From Genomes to Drugs. 2 vols. Weinheim, Germany: Wiley-VCH, 2002. Pp. 442+206. $320.

THIS REVIEW COVERS TWO BOOKS devoted to the state of the art in bioinformatics: a one-volume work, Current Topics in Computation Molecular Biology, and a two-volume work, Bioinformatics. Both are organized collections of articles by experts in the subfields of this discipline. There is substantial overlap between the two (two of the authors appear in both), but they are largely independent, and both present valuable perspectives. Current Topics tilts a bit more in the math direction (it has a chapter on DNA data compression), while Bioinformatics tilts a bit more in the medical direction (it emphasizes drug discovery, as the title indicates). Both will be useful for anyone wanting to know the state of knowledge in current areas of research in bioinformatics.

Although these books are not textbooks, they could be valuable references for courses in bioinformatics, especially because they include many recent topics not adequately covered in textbooks. Moreover, they include helpful introductory material for subjects such as Bayesian statistical analysis (in Current Topics), and the basics of molecular medicine and the sequence structure of eukaryotic genes (in Bioinformatics). In a rapidly developing field like bioinformatics, books like these are especially appreciated.

Many people rightly complain that the term bioinformatics does not have a uniform usage: its meaning ranges loosely from "genomics" to include all of computational biology. In the foreword to Bioinformatics, for example, bioinformatics is equated with computational biology, but the latter might include anything in biology involving computing, e.g., computational ecology. Despite the border disputes, the core of bioinformatics is certainly the genome and related biology. We have entered a new era, often called "post-genomic," but better described as "genomic-enabled." Much of bioinformatics will be devoted to frisking the genomic information we have now, to refining and extending it to more species, and to exploring further relationships among genomes. Both of these books include material on a variety of topics in genomic analysis. Both cover (multiple) sequence alignment and gene and promoter recognition. Current Topics also discusses genome rearrangement, phylogenetics, and sequence compression, and it includes a foreword by Temple Smith on the "challenges facing genome informatics."

The data-mining topics in Current Topics deserve special mention. To begin with, there is a separate chapter that provides an introduction to the subject and gives examples spanning from the genome to medical records. This chapter illustrates how broad the subject of bio(medical)informatics can be. Data-mining is in part pattern discovery, and thus work on finding complex signals for common diseases is an example of data-mining (Cox et al. 1999). Although both books have chapters devoted to genetic variations and linkage analysis (Pritchard and Przeworski 2001), automatic discovery of genes and promoter sites is also a type of pattern recognition problem. Current Topics organizes all of these topics in a separate section so that these connections are clear; it also includes a chapter on pattern recognition (clustering) for gene-expression array data. The latter is covered briefly in Bioinformatics in the context of time-dependent gene expression.

Beyond genomics, there are many new challenges ahead as we seek to understand the proteins that genes encode, how they are regulated, how they are modified after transcription, how they aggregate, how protein assemblies are involved in cell formation, how proteins interact with cells, and perhaps most importantly, the role of protein dynamics in disease (Fernández 2002). Thus proteomics will be a much larger endeavor than genomics has been in bioinformatics. The material on proteomics in Current Topics is limited to four chapters on the problem of protein structure prediction, or protein folding, and one chapter on docking of proteins. However, information in Current Topics on genomic databases and data-mining will be useful background for related...