I was disapointed in this book. In part that is my own fault as the authors have a different audience and a different approach to the one I was looking for. In part, however, I think O'Reilly has got it wrong by accepting the growing importance of Bioinformatics in the 'post-genomic' era justifies their publishing a book in this area, and then allowing its focus to be on yesterday's paper-and-string-and-Perl culture.
The book makes clear that it is aimed at the experimental biological scientist who wants to do his own bioinformatics on the cheap, and focuses on setting up Linux, finding free software on the web, and using Perl to automate it. There is a wide coverage of topics and programs, but everything except the Linux stuff is ultimately incomprehensible on its own. One would need another book to do anything with Perl, Shell scripting, SQL or to understand the sequence comparison or tree building algorithms, or microarray technology (no colour picture!), to name a few topics that are touched on in an unsatisfactory manner. The range of programs related to protein structure is greater than one finds in other bioinformatics books, but I imagine anyone in that area is au fait with this sort of thing and probably has specific software available.
Although there will be readers for whom this approach is suitable, I think it is important to criticise the philosophy behind the book. Although one can argue that Bioinformatics is part of computational biology, which has been around for a long time, it would be naive to recognize that it is current prominence is solely because of the human and other genome projects, and that its application is to a biological science performed on a different scale than in the past. I doubt that there will be many do-it-yourself biologists glueing together bug-ridden Perl sphaghetti code. The budgets will allow the use of commercial software or the employment of professional bioinformaticians to set up distributed systems using modern robust programming languages (i.e. Java - why are the authors scared of Java? Why does the word 'applet' not figure in the book's index) and proper commercial databases. This will be even more the case in industry.
The biological scientist has a need to understand the theory behind the programs he uses, and this is not provided here. The big audience for a good book on bioinformatics (which I have not yet found) is masters-level university students training in this area. Most of these will be employed by pharma or in supporting big research labs. Certainly they are going to be using Unix, but they will be using sequence software such as GCG, and learning SQL on Oracle databases, and being trained in Java. They need a book that explains the algorithms behind the programs, describes modern techniques such as microarrays properly and provides an introduction to unix and scripting languages that sees them in their appropriate context. Most biological scientists will want to make their bioinformatics queries from their desktop pc or mac in the context of a well designed GUI, even though the bioinformaticist who is concerned with serving these requests will probably be using unix machine.
Sorry to be so negative. The authors wrote about what they believed in, but...
|
