Flutterby™! : Developing Bioinformatics Computer Skills 2001-06-11 01:49:36+00

Developing Bioinformatics Computer Skills

2001-06-11 01:49:36+00 by Dan Lyke 3 comments

Finished leafing through the first O'Reilly offering to hit the bioinformatics wave, Developing Bioinformatics Computer Skills. I am not its target audience. It's a quick and overview of a lot of different tools for bioinformatics (although none in the area that I'm currently exploring), with a bit of the chemistry bundled in, but I get the feeling that it would really rather be two fairly high level books, one covering the chemistry and genomic biology, the other covering the tools. Or at least I'd rather it were that, 'cause then I wouldn't have had to skim over the tools bits to find the parts that were up my alley. Anyway, can't recommend it unreservedly unless you're a working biologist who knows all this stuff, and then my guess is that you're probably steeped in the culture of your profession enough that you already know most of it.

I'm about half-way through it; need to get cracking and finish it off so I can write a real review, but so far I'm less than totally impressed.

As Dan says, the book is unsure about who it's talking to, and consequently ends up covering too much basic material for on both sides. I think three chapters on building a Linux workstation is way too much -- what's wrong with saying "See $ORA_BOOK for more"? Finally, there are a large number of typos (frex, check out the DNA and RNA binding interactions in chapter 4 [ I think it was Chapter 4...]).

PS: Don't miss "all your bioinformatics worms are belong to us!" notice on the copyright page -- but ORA has said that was a mistake and will be coming out

PPS: Hey, Dan, are you doing anything interesting that you can talk about in public? (Or in private...)

Nothing cosmic, Phil's written some very cool code to do expression analysis, and I've been diddling with a Windows port of that in the hopes that we can turn it into a low-end product to get noticed to get funding to build a better version (already most genomics folks who see it agree that this system has a better interface and paradigm for working than the other software on the market), but now that we're actually getting data that hasn't been sanitized and cleaned up we're spending quite a bit of time trying to puzzle through what information from the Affymetrix machines really means.

Actually, once we get something that's in a state that it can be run on a wider format of data-sets and installed on end-user machines I was kinda hoping that I'd be able to prevail upon you to give me a view of how the software stacks up independent from what I hear from the context of the other folks involved in this endeavor.

So howzaboutit? You doing anything with expression analysis? Be interested in diddling with some software that seems to be fairly advanced in its visualization system relative to other packages, that has most of the clustering algorithms, and a plug-in architecture for building your own?

As far as having a look at what you've got, I'd love to. I've seen GeneSpring at work (haven't played with it), and I've seen a few hand-rolled interfaces, but nothing really stunning. If you've got something that deals with the data overload issue, that would be fantastically cool.