Counting fully on DNA

As a molecular biologist studying aspects of ecology, systematics and ecophysiology by analysing DNA sequences, Pavel A. Pevzner's book took me by surprise. I never appreciated how the growth of molecular biology was so constrained by the development of computational analysis. This book begins with one such example, tracking down the gene that causes cystic fibrosis, and demonstrates how the development of computational tools for genetic mapping were so crucial in studying this disease. In another example, "sequencing by hybridisation" using micro-arrays, it is shown that development of new technology in molecular biology can be seriously inhibited by computational problems, especially when experimental error is introduced into the "equation". Many other examples are provided that not only help to illustrate the principles involved in computational molecular biology, but also make this book fascinating.

Computational Molecular Biology , however, is not for the mathematically faint-hearted. To appreciate fully the algorithmic approach used in the computational problems presented requires a standard of training equivalent to A-level or even degree-level mathematics or computing. In this respect, the book is aimed at undergraduate or postgraduate students in bioinformatics or computational biology, or at professional scientists requiring further background in these subjects. Many biology students will struggle with this material, especially as parts of the text require background knowledge of mathematical references to appreciate properly the algorithms and combinatorial problems presented. Pevzner's book avoids many aspects of bioinformatics, such as phylogenetics, that are well covered elsewhere. However, the book explores many themes of relevance to modern genomics and proteomics. These include gene-mapping, sequence alignment, detection of signals and genes in DNA sequence, genome rearrangement and identification of amino acids and proteins from mass spectrometry. I found the chapters on detection of genes in sequence data and on amino acid identification especially useful. Gene identification from DNA sequence data is particularly difficult in eukaryotes because genes are split into small pieces (exons) interrupted by non-coding DNA (introns). Pevzner outlines many of the approaches used in gene prediction, including identification of start/stop signals in DNA (codon usage), ExonPCR and methods that involve DNA database interrogation.

The chapter on computational proteomics examines how amino acids are identified from the spectral graphs produced in tandem mass spectrometry. Given the present emphasis on proteomics in molecular biology, this chapter is a useful addition.

Alex David Rogers is principal investigator of Antarctic biodiversity, British Antarctic Survey, Cambridge.