Customer Reviews

Bioinformatics: An Introduction (Computational Biology)

5 star:		(0)
4 star:		(0)
3 star:		(1)
2 star:		(0)
1 star:		(0)

 

 

To those who (like me) think of bioinformatics as a set of tools used to study large-scale genetic and proteomic data, this book feels a little bit out of place. Its emphasis is on big-picture theory -- specifically, the rôle played by information theory in biology. It is not very unified -- for example the extensive theoretical ideas developed in the first part rarely make an appearance in part III, "applications."

The flavor of the book is perhaps best summed up by the following comparison. In Ramsden we find:

A) ~80 pages on information theory, probability, and complexity.
B) ~25 pages pages on sequence databases, alignment algorithms, and phylogeny.
C) ~5 pages on protein folding/structure/function prediction and drug discovery.

Compare to Lesk's Introduction to Bioinformatics:

A) ~8 pages.
B) ~129 pages.
C) ~90 pages.

Ramsden's approach reminds me of high-level, grandioise complexity science books (ex. Érdi's Complexity Explained or Mitchell's Complexity: A Guided Tour). To Ramsden, bioinformatics is not a set of tools but rather the more general study of "deviations from randomness" in the genome.

This sort of discussion can be very enlightening, and is important for developing a personal research world view. His discussions of information theory, for instance, helped give me a wider appreciation for the meaning and importance it has in science at large.

However, his big-picture discussion left precision by the wayside, and it wasn't always possible to determine what he meant as he hopped fleetingly from discussing the definitions of microstates/macrostates/KL-divergence to vaguely referring to the Shannon Index as "remembered information" and its quantity as the "intensity of selection," and later discussing the "durability of information." All these analogies are swell, but they are not defined with any degree of precision, and so the possible interpretations of what the author means are underdetermined by the text.

I suppose that is to say that the book, while chalked full of mathematical equations and concepts, is at times difficult to follow for its lack of rigor. More than once I wrote "WTF?" in the margin.

This does not mean it is worthless, and many researchers who have a sharp interest in unifying theory can probably find value in it while developing their own vista on what biological information should be and how it relates to probability theory.

While sections of this book are good and stand-alone introductions to basic concepts of information theory, molecular biology, and bioinformatics tools, this book is not a practical introduction to the tools and methods of bioinformatics as it is practiced on a regular basis. For that, read Lesk, or for a more mathematically rigorous (if outdated) treatment, see Waterman's (Introduction to Computational Biology: Maps, Sequences and Genomes.