Meeting Minutes 3:Stanford
Implementation Meeting 14 January, 2008
David Hill, Tanya Berardini, Chris Mungall all at TAIR (Carnegie)
Tried to pull out all 'biological qualities' from the regulation of BQ terms in order to see if they map to existing PATO terms. Initial mapping attempts were unsuccessful.
Looked at Chris' error report:
/scratch/regulation-report.txt
UNEXPECTED: finding terms that are children of ROBQ for which the BQ is not expected to be found in GO but the term really does exist in GO, the question then is, should they be is_a children of ROBQ or not?
OK: ignore
HIERARCHY: term has both ROMF and ROBP as is-a parents
will also detect cases where just one is_a parent to the regulation upper term but it is the wrong kind (i.e. a reg of MF sub-term that only traces up to ROBP)
MISSING_LINK: terms that should be under biological regulation via a direct is_a path but are not
NP: no parse, this category probably includes terms that D and T have already looked at and decided were ok types of a legitimate regulates parent, Chris will modify this code to discriminate between the ones already checked and those that haven't been.
Working with the following file:
scratch/gene_ontology_xp_regulates-pre_reasoned-no-intersections.obo
reasoner run on Chris' machine in batch mode, then saved, then had intersection tags removed
rough outline of steps:
loaded in OBOEdit 2.0
ran reasoner
Edit -> assert implied links
Reviewed and accepted the removal of redundant links.
Save in OBO 1.2 format.
substitute part_of for regulates/positively regulates/negatively regulates (switch back to part_of since regulates is not yet live to the community)
remove interontology links from process to function (like regulation of MF is part_of MF)
ACTION ITEM: (D and T, to be reviewed by Jen and Midori)
Create email for announcement of 'regulates' relationship implementation.
ACTION ITEM: update documentation for regulation terms, regulates relationships
Draft letter, regulates announcement
Subject: new relationship types in GO
Dear GO Community:
We are pleased to announce that The Gene Ontology Consortium will introduce three new relationship types -- regulates, negatively_regulates and positively_regulates -- into the Biological Process ontology.
For some time now, the GO content developers have been aware that regulatory processes are not necessarily integral to the processes (as previously indicated by the use of the 'part_of' relation) that they regulate. Nevertheless, regulatory processes have been represented as part_of the processes they regulate. We have long intended to replace these part_of relationships with a new relationship type called 'regulates'. We are now in a position to make this replacement.
Ontology developers have thoroughly reviewed the relationships involving regulatory processes and their targets to ensure internal consistency. If a term 'regulation of process X' exists in the ontology, it must be a valid subtype of 'regulation of biological process', and must have a part_of relationship (which will be transformed into one of the 'regulates' relationships) with 'process X' or be a valid subtype of another regulatory process.
We have also introduced two new high-level regulation terms, 'regulation of molecular function' (for example, 'regulation of cyclase activity') and 'regulation of biological quality' (for example, 'regulation of blood pressure'), to represent processes that regulate the activity of gene products and processes that regulate measurable biological attributes, respectively. Regulation of molecular function terms have been aligned with the corresponding terms in the molecular function ontology, and we are working to align regulation of biological quality terms with PATO.
WE PLAN TO REPLACE THE CURRENT 'part_of' RELATIONSHIPS IN THE ONTOLOGY WITH THE NEW 'regulates' RELATIONSHIPS ON MARCH 25,2008.
Concurrent with this replacement, we will also introduce new 'intersection_of' tags into the OBO stanzas of the regulation terms; these tags can be safely ignored by software with no negative consequences.
Software developers should ensure that their procedures for loading the ontologies into their resources are compatible with these changes by this date. Below is an example of a current OBO 1.2 stanza for a regulates term and a new stanza with the regulates relationship in place.
CURRENT STANZA
[Term]
id: GO:0000019
name: regulation of mitotic recombination
namespace: biological_process
def: "Any process that modulates the frequency, rate or extent of DNA recombination during mitosis." [GOC:go_curators]
narrow_synonym: "regulation of recombination within rDNA repeats" []
is_a: GO:0000018 ! regulation of DNA recombination
relationship: part_of GO:0006312 ! mitotic recombination
NEW STANZA
[Term]
id: GO:0000019
name: regulation of mitotic recombination
namespace: biological_process
def: "Any process that modulates the frequency, rate or extent of DNA recombination during mitosis." [GOC:go_curators]
synonym: "regulation of recombination within rDNA repeats" NARROW []
is_a: GO:0000018 ! regulation of DNA recombination
intersection_of: GO:0065007 ! biological regulation
intersection_of: regulates GO:0006312 ! mitotic recombination
relationship: regulates GO:0006312 ! mitotic recombination
We strongly suggest that you test your loading scripts to ensure that your database doesn't go down in flames. There is a test OBO 1.2 file available at:
go/scratch/gene_ontology_with_regulates_relations_test.obo
Please feel free to contact us with any questions or concerns about this change.
Sincerely,
The GOC regulates working group
Meeting Minutes, Day 2, 15 January 2008
David Hill, Tanya Berardini, Chris Mungall at TAIR (Carnegie)
Reviewed diff between regulation-nugo-fixed.obo and current gene_ontology_edit.obo. All proposed new relationships are ok. Confirmed that redundant links have been removed.
Put up draft letter for 'regulates' premiere for review.
Skype conversation with Midori and Jen to bring them up-to-date on yesterday's progress.
Removed TypeDefs for 'regulates' relationships from .obo file before committing changes.
Chris reran analysis and file processing scripts on latest gene_ontology_edit.obo file and we committed this to the repository after sanity check. Revision 5.647.
Stats: 969 new links 302 deleted links
Worked through Chris' QC report after the rerun on new, committed file:
new QC report name: /scratch/regulation-report.txt
UNEXPECTED:10 (DONE)
MISSING LINK:7 (DONE)
HIERARCHY:28 (DONE)
NO_TARGET: almost all ROBQ and ROMF children or ones that were ok subchildren of higher level ROBPs: 383 (STARTED)
NP (NO PARSE): appear to be subclasses of 'regulation' terms that are not named 'regulation of %':1091
OK:3080
Issues with new file:
Regulation of receptor activity has now inherited a regulation of some type of binding because receptor activity is_a binding. However the regulation of the receptor is not always a regulation of the binding, so we needed to go through and delete those.
Other discussions:
We talked about the relationships between function and process due to the ability of a regulates term to be both a RoMF and RoBP. We decided we will need both part_of and has_part relationships for the MF to BP links
Second error report: scratch/regulation-unimplied-links.txt
These represent cases where the regulates hierarchy doesn't match the process hierarchy.