Bioinformatic book review: It came from the lab!
- By Jack Vaughan
The discovery 50 years ago of the DNA double helix lagged the invention of the digital computer by only a few years. It took a number of years thereafter for computers and genetic research to become entwined but, in recent years, the blend of the two disciplines has produced some interesting results, not the least of which is the nearly full model of the human genome.
In recent years, the field of bioinformatics has arisen as a somewhat distinct application area in computing. Spurred by the push to map the human genome, distributed hardware architectures have been re-energized. At the same time, XML databases have found another area where they may excel vs. relational kinds of databases. Also, Perlsters and other scripters have found plenty to do by gluing odd data types and tasks together.
Programmers who have an interest in this new field are at something of a disadvantage when it comes to surveying the territory. The Web is a font of information, but it is often disconnected information without much context. Once again, books may be the best way to get up to speed, but finding the right one is not easy. They range from very wide, very expensive and very dense text books that are an amalgam of learned papers and require more than a journeyman's mastery of biochemistry, to narrower, highly focused computer software books that hardly touch on biology at all.
Enter Bryan P. Bergeron, M.D., of Harvard Medical School and M.I.T. His "Bioinformatics Computing" (Prentice Hall PTR, 2002) makes a genuine attempt to span the fields of biology and computing, while still paying each field its due.
That isn't easy. This reader sorely needed grounding in what a microarray is, just to name one topic. A reminder on the basic types of data schema was welcome. But a discussion of network topologies, though numbering just a few pages, seemed interminable.
In the end, there is not that much detail about scripting techniques used to stage the complex (and sometimes deceptively simple) data types of genetic and related research. But that type of information is handled in other books.
Math search filters are also described in passing. Did you know there is such a thing as a Hidden Markov Model -- I did, but I'd forgotten -- and that they can be applied to DNA data? Good to know, but clearly you have to move on to other texts to begin to build and apply such models, assuming you have prepared your data correctly.
This is niggling. Bergeron's book is a worthwhile piece of work that is determined to inform readers. Technologists will want primers on these subjects, and they could hardly do better than Bergeron. There is every reason to think that "Bioinformatics Computing" could serve any lab where a translation book is needed so that the genetic hacks and computer geeks can work together and talk something approximating the same language.
For other Programmers Report articles, please go to www.adtmag.com/article.asp?id=6265
Jack Vaughan is former Editor-at-Large at Application Development Trends magazine.