XP:biological process xp self: Difference between revisions
No edit summary |
|||
Line 106: | Line 106: | ||
* "X by Y" => "X by Y" is_a X AND "X by Y" is_a Y | * "X by Y" => "X by Y" is_a X AND "X by Y" is_a Y | ||
** For details of an example, "regulation of X by regulation of transcription", see [https://sourceforge.net/tracker/?func=detail&aid=3089689&group_id=36855&atid=440764 SF 3089689] and [http://fafner.stanford.edu/pipermail/ontology-editors/2010-November/001229.html this email thread]. | |||
* "X via Y" => "X via Y" is_a X AND "X via Y" has_part Y | * "X via Y" => "X via Y" is_a X AND "X via Y" has_part Y | ||
* "X involved in Y" => "X involved in Y" is_a X AND "X involved in Y" part_of Y | * "X involved in Y" => "X involved in Y" is_a X AND "X involved in Y" part_of Y |
Revision as of 10:40, 16 November 2010
cross-products within BP
Implementation responsibility: Jennifer Deegan
Status: part_of xp defs are in the main ontology
biological process xp self Cross Product File Metrics
Progress notes: biological_process_xp_self_ProgressNotes
This is mostly temporal xps: definitions of terms based on a genus term and a larger containing process
This extends:
And should thus be non-redundant with them; you should see no xp defs involving the regulates link.
However, we do expect xps such as the following:
[Term] id: GO:0006933 ! negative regulation of cell adhesion involved in substrate-bound cell migration intersection_of: GO:0007162 ! negative regulation of cell adhesion intersection_of: part_of GO:0006929 ! substrate-bound cell migration
this is technically not a regulation xp: we can further expand the xp using the regulation xps to get
( (biological_regulation ^ regulates(cell_adhesion)) ^ part_of(substrate-bound cell migration) )
Methods
Obol was used to generate the initial version of the file. Obol grammar: obol_biological_process_xp_self prolog grammar.
In contrast to many of the other XP files, an additional approach was used to seed potential xps defs: for each term, is_a and part_of pairs were tested to see if they were unique to that term. If unique, the links were used in the xp def. Note that this inevitably leads to "false positives" - necessary and sufficient conditions that are also satisfied by other terms. These were all examined by cjm, but some may have slipped through.
ontologies used
- GO (BP)
- RO
- ro_proposed
availability
- "neural plate elongation" [GO:0014022] is_a "embryonic morphogenesis" [GO:0048598]
The full obol explanation is as follows:
"neural plate elongation" [GO:0014022] is_a "embryonic morphogenesis" [GO:0048598] because of rule: genus_differentia : "embryonic morphogenesis" [GO:0048598] is defined as { a "anatomical structure morphogenesis" [GO:0009653] that part_of "embryonic development" [GO:0009790] } , and it is the case that "neural plate elongation" [GO:0014022] is_a "anatomical structure morphogenesis" [GO:0009653] and "neural plate elongation" [GO:0014022] part_of "embryonic development" [GO:0009790]
Which isn't exactly normal english.
Loading the xps into OE, we can ask for an explanation here (implied links are blue dashed lines):
It's still not exactly clear...
Here is my explanation:
- embryonic morphogenesis = any morphogenesis THAT is part_of embryonic development
- neural plate elongation is part_of embryonic development (by transitivity - can be seen in the graph above)
- neural plate elongation is_a anatomical structure morphogenesis (this is asserted in the ontology by editors)
Problems
Mitosis etc
obol parses 'mitotic spindle elongation' as
- spindle elongation THAT part_of mitosis
i.e.
[Term] id: GO:0000022 ! mitotic spindle elongation intersection_of: GO:0051231 ! spindle elongation intersection_of: part_of GO:0007067 ! mitosis
Perhaps these should be manually fixed to
- spindle elongation THAT part_of mitotic cell cycle
I have to confess I find this part of the ontology a bit confusing
part_of vs during / involved in
obol rules:
- "X by Y" => "X by Y" is_a X AND "X by Y" is_a Y
- For details of an example, "regulation of X by regulation of transcription", see SF 3089689 and this email thread.
- "X via Y" => "X via Y" is_a X AND "X via Y" has_part Y
- "X involved in Y" => "X involved in Y" is_a X AND "X involved in Y" part_of Y
- "X during Y" => (temporal distinction) "X during Y" is_a X
- really no part_of or other relationship between "X during Y" and Y