[Ontology-editors] function-process links downloads

[Chris Mungall]

Just a reminder that all the various curated and auto-generated links  
generated as part of this project are available in the scratch dir in  
cvs:

	http://www.geneontology.org/scratch/fp-links/

Comments welcome.

There's quite a few, so I have added a README file. Let me know if  
anything needs explained further.

--


All obo files are designed to be loaded in conjunction with GO

Alternatively they can be examined in a text editor

For more information, see:

   http://wiki.geneontology.org/index.php/Category:Function-Process

== Curated ==

The following links were curated manually. They are all has_part links

* glycolysis-p2f.obo (Harold)

* TCA-cycle-p2f.obo (Harold)

* FXP-2.obo (Harold) - collection of various processes

* electron-transport-p2f.obo (ET group)

   Note that these links require the new terms in electron-transport- 
newterms.obo

== Single-step part_of links ==

For single-step processes we can make direct part_of links between F
and P - gene products can be propagated across these

* single_step_fp.obo

   Based on biological_process_xp_chebi and  
molecular_function_xp_chebi. If the xp defs are the same, make a link

* kinase_phosphorylation_fp.obo

   Based on text matching

* transporter_transport_fp.obo

   Based on text matching

== Reactome ==

The following were mined from the biopax reactome dumps

* reactome-all.txt

   table showing reactome P-F links. Taken from reactome-biopax.
   If a reactome reaction corresponding to a MF is executed in a  
reactome pathway corresponding to a BP, there
   should be an entry in the table.
   Uses the reactome to GO mappings included in the reactome biopax  
exports.
   The final column shows all species for which this P2F link exists,  
together with how the P is named in reactome

* reactome-all.obo

   above, converted to .obo format. Can be loaded in OE at the same  
time as GO

* reactome-all-imports.obo

   point OE at this; will load both GO and PF links

* reactome-*.p2f

   individual P2F links on a per-species basis.

== Mining Interpro ==

Can we mine f->p and/or p->f links from interpro? If one domain
implies both a function and a process, it also implies a direct
correlation between the F and the P

* p2f-from-ipr.obo

* f2p-from-ipr.obo

== Cellular component to F/P ==

Not technically F-P, but included here for completeness:

* component_function.obo

* component_process.obo