nodalpoint.org

I'm skimming through the latest TOC of the journal Bioinformatics and making a few observations.

First, it seems to have become devoted almost entirely to methodology, algorithms etc., rather than biological discovery. That's OK - there's a place for that. I'd submit to Bioinformatics if I had a small, cute discovery made with minimal effort, otherwise I'd be thinking Genome Research.

A couple of papers caught my eye for various reasons. Goldovsky et al. on the CoGenT++ environment is interesting - I've always liked the work of the Ouzounis group. What this tells me is that dynamic web-based databases are now pointless bioinformatics Honours projects - the professionals do it so much better. Let's teach them the basics of SQL and CGI and leave it at that.

Crooks et al. on pairwise alignment got my attention, because it's 6 pages and concludes with a negative result - namely that building covariation of neigbouring residues into pairwise alignment algorithms doesn't make them any better. I read this partly to improve my understanding of algorithm design, but it got me thinking about whether negative results should be reported more often. I think that they should.

Mao et al. on genome annotation using KEGG is also worth a look for the free downloads, though it seems to be largely a Python implementation of PRIAM. I like the PRIAM people - I found a bug in their Perl/GD code earlier this year and they were very prompt in fixing it.

Aside from that it seems a smaller issue than usual and they've undergone some website redesign which I think is not quite complete.