[BioRuby] Programmatic interfaces to online-only bioinformatic tools

[Ben Woodcroft]

Hi there,

I'm after an opinion on best practice for accessing those bioinformatic
tools that are only accessible via web-interfaces, without any way of
downloading the program itself. As I often want to throw an organisms's
whole proteome at the tool (in my case ~5500 proteins), doing it one at a
time is obviously not feasible. Do people think that writing a wrapper that
hammers the website in question is a nice thing to do, provided I wait 1
second in between each upload (1 protein/upload)? From what I can
understand, processing each protein is not computationally demanding and
results are returned almost instantaneously when querying the server
manually, unlike a blast server.

Some background:
I'm really like this bioruby plugin thing - thanks to those involved in its
creation. I've written a bioruby plugin
https://github.com/wwood/bioruby-cnls_screenscraper which queries
http://nls-mapper.iab.keio.ac.jp/cgi-bin/NLS_Mapper_form.cgi that does what
I've described above. Not quite as well documented as it might be, but it
appears to work and isn't too complicated.

Thanks,
ben

-- 
Ben J Woodcroft, BE (Hons)

PhD Candidate
Ralph Laboratory
The University of Melbourne
Melbourne, Australia