Map PDB match coords to Pfam protein domain

Question

[Anand Rao]

I am using hhpred to find matches to my peptide queries (45-50aa long) against the PDB database.

For the matching coords in the top-ranked PDB file, I want to check whether those coords correspond fully or partially to my (Pfam-defined) protein domain of interest.

How can I perform this mapping of PDB match to Pfam domain name in a high-throughput manner at the command line?

I realize there are links to 3rd party tools for parsing the PDB files ( http://bit.ly/1JcsSt6 ), but I've never played around with PDB files, so I thought I might first glean ideas from the helpful forum members here. If you think I could ask this question in another forum more qualified to answer this question, I would appreciate that as well. Thank you!

Answer 1

[peter.rose]

You can retrieve PFAM assignments using our tabular report RESTful web services.

We run PFAM HMMs to update PFAM assignments.

Example URL with two PDB IDs:

http://www.rcsb.org/pdb/rest/customReport?pdbids=1A0K,1AA2&customReportColumns=pfamAccession,pfamId,pfamDescription&format=csv&service=wsfile

This produces the following .csv file:

structureId,chainId,pfamAccession,pfamId,pfamDescription
"1A0K","A","PF00235 [2-126]","Profilin","Profilin"
"1AA2","A","PF00307 [5-106]","CH","Calponin homology (CH) domain"

If the list of PDB IDs becomes long, you need to use POST rather than GET in the web service call.

There are links to code examples at the bottom of this page.

Answer 2

[andreas.prlic]

RCSB PDB is scripting the HMMER web serives on a weekly basis and in addiiton to the web-service link posted above, the data is also available from here.

Answer 3

[Anand Rao]

I may have found my answer at http://www.ebi.ac.uk/pdbe/docs/sifts/quick.html

I must cross reference an earlier question here at biostars that elicited numerous answers: Merging Data From Pfam Pdb And Uniprot

I am not sure whether the mapping information is most up to date, but its a really good start for me in any case.