Let's say I have a set of UCSC names of genes.
Is there a database (or databases) I can query through a RESTful request (or something else I can automate on the command line), that would tell me what protein domains are contained in translated genes, such that I could categorize them?
I don't know much about protein databases, so I apologize in advance for what will likely be a n00b question. Thanks for your help.
•
written
by
Alex Reynolds
You could use NCBI's RPS/PSI blast against their Conserved Domain Database (CDD) database. I don't believe they have a restful interface, but they do have a web/CGI interface. In addition you are free to download the entire CDD and run your RPS blast locally. This is what I would recommend that you do.
However, they do have a sample Perl script available that demonstrates how to automate the search.
They have a collection of internal (CD) data along with TIGR, Pfam, COGs, and more.
Thanks, I'll investigate the CGI interface. I can send forms programmatically with curl, wget, etc.
You could also try the UniProt REST interface: http://www.uniprot.org/faq/28
You might use the gff format, e.g. for human gene alarp4b: http://www.uniprot.org/uniprot/?query=gene%3alarp4b+reviewed%3ayes+organism%3a9606&format=gff
Log in to answer this question.
An easiest way would be to use Biomart. When you select for Attributes, you get an option of choosing information related to Protein Domains and Families. You can then download domain related information from there. They have PFAM, Interpro, TIGRFam data.