> I'm toying with the idea of doing a presentation on bioinformatics at > YAPC. Partially this is motivated as a selling point to my boss(es), as > we have a teaching/community outreach requirement for one of our grants > I might be able to squeak this in under.
A few years ago, I made some extensive modifications to an application called SeqWeb, that is basically an extensive web interface to a handful of small BLAST apps (GCGBlast). I found it both challenging and interesting. The primary reasons for the modifications were that the scientists wished to use FASTA format input rather than GCG, WU- rather than GCG-BLAST, and wanted to offload all of the jobs onto LSF load-sharing clusters. SeqWeb performs some fairly elaborate parsing, both on the input, and the output. It also monitors a forked process, and can maintain the browser state until the BLAST is finished, which can be an issue where LSF just returns immediately after posting the job. I suggest that you could break-down a project like this into it's component parts, namely complex regexes, and monitoring third-party processes or remote execution. I should think that you could take some of your experiences and easily create a similar talk (perhaps a lightning talk). Just a suggestion, Grant M. P.S.> Are they looking for any help over there at Harvard? ;-).
