Hi,

Does anyone know a convenient way to identify membrane bound genes given a list of hundreds of genes?

AP

You won't find any region of similarity between all membrane bound proteins, but there are similarities between groups. Look up Pleckstrin homology for example - proteins sharing a similar mechanism of binding lipid anchors:
http://www.jbc.org/cgi/content/full/275/20/14873

Are you differenciating between membrane bound and transmembrane? Transmembrane proteins are easier to identify from sequence.

The answer to your question could involve literature mining - what are you starting from - is your list a list of names, sequences, fragments of sequences...?

There are a number of tools for predicting protein subcellular localization out there that can help. Many of these are linked to from here

http://www.psort.org/

bgood said:

There are a number of tools for predicting protein subcellular localization out there that can help. Many of these are linked to from here http://www.psort.org/

Yes, psort would be a good solution for eubacteria, though the eukaryota version is quite old. I'm actually unclear at the moment whether you want to find "known" membrane-localised proteins or predict localization (which is what psort and others do). One is a text-mining problem while the other is a classification problem. If you want to predict localization based on signal peptides in the protein sequence, try signalP or SubLoc.
As for text-mining the known ones, I can't help you there.
Cheers.
Ryan

Interesting that you should try to seperate text-mining from classification. PSORT-b's main competition comes from a program called Proteome Analyst that builds its classifier by integrating Blast and and text/tag based classifiers. Pretty cool really.

http://www.cs.ualberta.ca/%7Ebioinfo/PA/Sub/index.html

There are links to lots of different predictors up on the PSORT website - I don't think their (PSORTs) new eukaryotic predictor is ready , but many of the others are.

I found that TMpred is a simple tool to use, but there are no doubt other ones out there.

You might want to check out this paper to see the tool in use:

Example of TMpred in use

Another transmembrane prediction tool developed by researchers at Harvard Medical School:

PLoS ONE Paper