Neural Networks and Genome Informatics

Rating:
Summary: Timely text for beginners and experts alike
Review: In contrast to so many other bioinformatics books this one is written by genuine experts who have first-hand experience in both computer science and modern biology.

Each chapter is a masterpiece of clarity and good judgement in selection of topics to be covered. The book contains a large glossary of terms which makes it accessible for multidisciplinary readers. The authors made considerable effort to provide unbiased selection of most appropriate references following each chapter. This makes the book a superb research monograph for the specialists in addition to being a suitable text for educated beginners.

The book should be read by computer scientists who contemplate doing work in bioinformatics as well as by biologists who contemplate working in bioinformatics. Anybody who wants to design neural networks for specific biological applications will benefit the most from reading this text. Anybody who just wants to understand how and why neural networks can be used in biology will benefit from reading this book as well. Practicing computational biologists and bioinformaticians should have this book available as a desk reference. Psychologists, cognitive and social scientists who are interested in neural networks and artificial learning will likely benefit from reading this book as well.

I hasten to add that it would be really good for the book and for its readers if the publisher considered either lowering the price or printing a cheaper paperback edition.

Rating:
Summary: Good as a literature survey
Review: This book serves well to introduce the reader to the literature on the applications of neural networks to bioinformatics. It falls short however in giving an in-depth view of how neural networks operate and does not include any source code. Performance issues with the use of neural networks in genome informatics should have been given a more careful treatment. Considering its price, this is disappointing. A reader could obtain the required reading material on this subject from an online search. An instructor in a course in bioinformatics might use this book as a reference source however. Those who have used neural networks in other fields might be able to use the book as a guide to applying them to genome informatics. Thus the book could be viewed as a (very expensive) literature review article, but it does include some interesting remarks at various places: 1. Amino acid groupings that are found automatically by a Kohonen self-organizing map. 2. Feature representation and input encoding. 3. The discussion on cross-validation. 4. The discussion on protein secondary structure prediction. Genetic algorithms are mentioned here, so readers not familiar with these will have to gain the background elsewhere.

Rating:
Summary: GreatBook
Review: This is a book that comes out at the right time, a time when tons of information from genomics and several improved analysis tools based on great ideas are both becoming available. I believe readers from a broad range of academic background will benefit from the integration of knowledges from genome informatics, statistics, computer science, engineering, and mathmatics, a feature that this book exemplifies.