Most Helpful Customer Reviews
19 of 19 people found the following review helpful:
5.0 out of 5 stars
How does sequence alignment actually work?, November 23, 2003
This review is from: Blast (Paperback)
If you want to understand the nuts and bolts of how sequence alignment works, then this is the book for you. It will be especially useful for BLAST users who want to understand how it actually works and also for developers who don't know much biology, struggle with the math, but have no problem reading a perl script. The book is basically divided into:
0. A Foreword by Stephen Altschul (the co-creator of BLAST)
1. A quick web intro to a BLAST search
2. Sequence alignment and how the algorithms work
3. Blast and how the Blast statistics are calculated
4. The different types of Blast e.g. WU-Blast
5. Approaches to Performance speedup
6. Reference sections on BLAST parameters The real key is that this book neatly splits the difference between academic texts and papers which are quite often too difficult to read without sufficient background (and they are not precise about the implementation anyway) and the user-manual type texts which don't discuss the theory at all. One of the best chapters (in my view) is chapter three, where they explain and illustrate the workings of the Needleman-Wunsch and Smith-Waterman algorithms for global and local alignment. If you read the text, then study and run the included perl code, you WILL understand how they work, but be prepared to spend several hours trying different examples. The real advantage of this approach is that you get a deep, practical understanding of how alignment actually works, that you just can't get from reading a mathematical treatment of the subject. Once you understand this chapter, you are actually sufficiently expert to get inside alignment code and modify it for your own purposes. Ian Korf does continually emphasize that the algorithms may look clever, but they are, in the end, robotic in that they will quite happily align complete rubbish if you are not careful about controlling the algorithm and thinking carefully about the results you get. There are a couple of mistakes in the diagrams (chap 3), that are addressed in the errata, but the perl code is correct. Finally, because this book is about BLAST, it doesn't mention other methods of sequence alignment such as Hidden-Markov Models or methods of multiple sequence alignment. Perhaps they'll do a book on those as well one day..
Help other customers find the most helpful reviews
Was this review helpful to you? Yes
No
14 of 16 people found the following review helpful:
5.0 out of 5 stars
Author comments, August 6, 2003
This review is from: Blast (Paperback)
As the first reviewer mentioned, the book is not a fast read. In order to run BLAST properly one must understand how and why it works. BLAST exists at the intersection of molecular biology, computer science, and statistics. This might sound intimidating, but once you read about these topics in chapters 2-4, you'll see that it isn't so complicated and it all fits together nicely. We know that BLAST users come from a variety of backgrounds and we have therefore written the book for a general audience. As a result, the book is more than just a BLAST manual, it's also a friendly introduction to computational molecular biology. Writing this book took a lot of time and effort. It went through some
painful transformations. The authors waged many battles against
themselves and each other to bring to you the kind of book we wished we
could have bought several years ago. We'll feel our work was justified
if you approach your next BLAST search as a scientific experiment and
not a Google search. And if we've helped some of you to embark on a new
career/hobby in bioinformatics, drop us a line, it's sure to make our
day.
Help other customers find the most helpful reviews
Was this review helpful to you? Yes
No
4 of 4 people found the following review helpful:
5.0 out of 5 stars
Blast User's Bible, January 28, 2006
This review is from: Blast (Paperback)
This is the place to start for anyone using NCBI BLAST. It's a thorough description of the various BLAST programs for nucleotides, amino acids, and codons. The book offers a biology refresher early on, but this is aimed mainly at people with serious interest in BLAST - people who normally won't need that. Next, it discusses traditional dynamic programming alsorithms for local and global alignment. Then, in just a few pages, it summarizes the mathematical meanings and derivations of the various BLAST scores (raw scores, P-values, ane E-values). The discussion just skims the theory, but will help the reader make sense of the programs' output. Those 75 pages set the background; the next 250 contain the real meat of the book. They cover the various BLAST programs, options, and outputs. More than that, these sections discuss setting up experiments based on BLAST, and how to deal with the problems you're likely to encounter. This could be a bit more explicit about how PSI_BLAST works (and why it sometimes doesn't), but coverage is generally strong. A few things are weak, like emphasis on the fact that experiments aren't strictly repeatable. For example, if you exactly replicate today's test next week, even if all of the other input is identical, you might still get different (and worse) E values, since they depend on the size of the database. PSSMs get little if any discussion. Also, details about internals are weak - but this is a user's book, not an implementor's, so that's a matter of scope rather than sufficiency. Most of the book's points are illustrated with actual output or with Perl code - the lingua franca of bioinformatics, for some reason. If you're serious about using BLAST and about understanding what it's really telling you, this is the book to own. //wiredweird
Help other customers find the most helpful reviews
Was this review helpful to you? Yes
No
|
