Misplaced Pages

Revision as of 20:26, 21 March 2011

Prosite vs PROSITE

Both are now accepted fields thanks to Boghog2. Abergabe (talk) 13:29, 24 June 2010 (UTC)

Suggested changes

Boghog suggested changes in this template here. I see three issues here.

1. Making PDB RCSB query. Can we actually make a query to identify all PDB structures that belong to PFAM family PF0XXXX? The suggested template does not allow it. If we can, that would be a significant improvement, and I strongly support it because such version would automatically update the list of PDB files that belong to each PFAM family! This should be possible because current PDB RCSB version provides PDB->PFAM mapping, but I do not know how to do it, especially in the template.
2. Links to list of PDB files in Pfam. That might be excessive because we have a link to PFAM already, but it does not hurt. Support.
3. Links to PDBsum entries. They are not present in the new version proposed by Boghog. If we can make a query as for PDB (see #1), we could indeed replace all current links by a query (this should be possible because PDBsum provides PDB-PFAM mapping). If not, let's keep current links to specified PDB files.Hodja Nasreddin (talk) 02:44, 10 March 2011 (UTC)

ad 1) I just added such a link to the RCSB source code. It will be available on the public site soon. Also sent a link to a test server to Boghog and I think the link will work fine for him. I'll post an update here with the details once this is available for the public. --Andreas (talk) 03:34, 10 March 2011 (UTC)

In response to Hodja three points:

1. Making PDB RCSB query. The sandbox version (see testcases) already returns all the structures that contain the Pfam family PF0XXXX. The sandbox version currently uses the {{Pfam2pdb}} template (which in turn is based on this pdb to pfam accession number list) to accomplish this. Because of the enormous size of this template, I asked Andreas if he could enable a Pfam query link to the RCSB PDB. As he mentioned above, he sent me a test link that I have verified works. As soon as a public version becomes available, we will include it in the {{Infobox protein family}} template.
2. Links to list of PDB files in Pfam The advantage of this link is that provides detailed information about the precise location of the Pfam domain within each structure. I know this information is also provided in some of the other links graphically, but this particular link provides a concise text summary of this information.
3. Links to PDBsum entries As can clearly be seen in the testcases, individual PDBsum links are included in the sandbox version. In summary, what is currently implemented in the sandbox are query links to (1) Pfam, (2) RCSB PDB, and (3) PDBe, and each of these links return all of the structures associated with a given Pfam domain. It would be nice if PDBsum could also provide such a link. In the mean time, the individual PDBsum links will still be displayed. Boghog (talk) 07:59, 10 March 2011 (UTC)

All right, everything sounds great. Let's make these modifications in template using link/query provided by Andreas. BTW, are Pfam-PDB mappings identical in Pfam and PDB databases? I had an impression that PDBe does such mapping independently, as soon as new PDB files are released. That's important because Pfam is normally updated once a year, but PDB is updated every week.Hodja Nasreddin (talk) 17:52, 10 March 2011 (UTC)

The {{Pfam2pdb}} and {{Pfam2PDBsum}} templates were created from the same pdb to pfam accession number list so they both should return identical lists of structures. These templates are currently up-to-date. Ideally I would like to replace both templates with direct query links to the external databases, but if these take a long time to implement, the templates can be updated from time to time. Boghog (talk) 22:12, 12 March 2011 (UTC)

Just to add, RCSB PDB loads PDBe-SIFTS files (which provide the mapping) as well as Pfam on a weekly basis --Andreas (talk) 22:53, 12 March 2011 (UTC)

 Done The new version of the {{Infobox protein family}} has now been put into production (diff). I made a new {{Pfam2PDBsum}} that is transcluded into the infobox and removed completely the PDB parameter (all external PDB links are now derived from the Pfam parameter and the PDB parameter has been deprecated). As soon as pfam query links to the RCSB PDB and PDBsum databases are available, these can replace the {{Pfam2pdb}} and {{Pfam2PDBsum}} templates respectively. Cheers. Boghog (talk) 16:27, 12 March 2011 (UTC)

This is serious improvement. I quickly tested it for PH domain. As expected, some of the most recently released PDB files now appear in PDB (and PDBsum), but not in Pfam link (e.g. 3pp2). However, something strange is happening with PDBe link . It searches for PF00104 (Ligand-binding domain of nuclear hormone receptor), in addition to PH domain (PF00169) and therefore retrieves a much larger number of files. This should be fixed. Otherwise, great work! Hodja Nasreddin (talk) 21:33, 12 March 2011 (UTC)

Thanks for catching the bug. Hopefully it is now fixed. Sorry about that. Boghog (talk) 21:56, 12 March 2011 (UTC)

Great! PDBe search provides a nice sortable table , but they forget to include a field with UniProt code (this should be done as in Pfam: ). But this is their problem.Hodja Nasreddin (talk) 22:09, 12 March 2011 (UTC)

Perhaps we should not make collapsible three links with "Available protein structures", but only collapse the list of PDBsum files. Another question: should we remove "OPM protein" and leave only "OPM family"? I think this is something for you to decide since you work so much with this template.Hodja Nasreddin (talk) 22:28, 12 March 2011 (UTC)

Concerning the OPM family/protein links and the collapsable views, I don't have a strong feeling one way or the other. Perhaps we should experiment with the sandbox version first and leave the production version as is for a few days to see if others express an opinion. If no one objects, I would be happy to change the production version. Boghog (talk) 22:41, 12 March 2011 (UTC)

Great work Boghog, thanks for this improvement! As promised, I will post an updated RCSB link here once it is available to the public. Currently ETA is early April. If you like customizeable tables, Hodja, check the "Generate Reports" drop down at RCSB... --Andreas (talk) 22:53, 12 March 2011 (UTC)

Agree. My personal suggestion would be to keep everything as it is right now. I really like this new version. Hodja Nasreddin (talk) 23:17, 12 March 2011 (UTC)

New PDBsum hooks

The good people at PDBsum have added support for queries of their site with Pfam accessions eg . These should be much more consistent with the other structure links. I hope these are of use here. --Paul (talk) 15:58, 21 March 2011 (UTC)

 Done I am very impressed with the quality of this new PDBsum layout. I have already included the new link in the template (see beta-lactamase for an example). I am open to suggestions for tweaking the display of the link (I wasn't quite sure what to call it). Thank you Paul Gardner, Alex Bateman, and especially Roman Laskowski for implementing the link and for producing this great looking report! These new pfam structure links greatly increase the value of the pfam infoboxes and at the same time, eliminate the need to update the links. A double win. Thanks again to all for your help. Cheers. Boghog (talk) 20:26, 21 March 2011 (UTC)