This is the follow-up sweep from the upper-level refactor #602. Every pre-coordinated value term in PATO (e.g. increased weight) should carry an EquivalentTo defining axiom of the form <parent> and has_modifier some increased, so that the compositional and pre-coordinated forms come out equivalent under reasoning.
The refactor in #602 (PR #603) introduced PATO:0102003 (increased) and PATO:0102004 (decreased) as comparative modifier characteristics, on the has_modifier (RO:0002573) property. Per the uPheno-call agreement of 2026-06-12, this is the direction we want: pre-coordinated qualities defined against the modifier terms rather than against the old increased/decreased grouping classes. I deliberately left the actual sweep out of the upper-level edit and parked it here so #602/#603 stayed focused.
Four cases were left half-done in §5 of #602: their is_a link to the old increased/decreased quality grouping classes (PATO:0002300 / PATO:0002301) was dropped, but no has_modifier axiom was added to replace it. They currently only carry their RO:0015007 / RO:0015008 intersection_of axioms. Those are:
- PATO:0001782 increased intensity
- PATO:0001783 decreased intensity
- PATO:0000470 increased amount
- PATO:0001997 decreased amount
Technical note worth flagging, because it changes what this sweep actually does: the RO:0015007 / RO:0015008 axioms are logically independent of has_modifier some increased/decreased. PATO:0102003 / PATO:0102004 have no equivalence axiom tying them back to RO:0015007 / RO:0015008, so adding the modifier axioms is a genuine semantic addition, not a redundant no-op. That means we cannot assume the sweep is safe by inspection — a reasoner / QC run is required afterwards to confirm nothing goes unsatisfiable.
Scope for this issue: enumerate the full set of pre-coordinated value terms, add the has_modifier some increased / has_modifier some decreased defining axioms, and run the reasoner. This should land before curators start composing phenotypes with the new modifier terms, otherwise the compositional and pre-coordinated forms will drift apart.
Related
This is the follow-up sweep from the upper-level refactor #602. Every pre-coordinated value term in PATO (e.g. increased weight) should carry an
EquivalentTodefining axiom of the form<parent> and has_modifier some increased, so that the compositional and pre-coordinated forms come out equivalent under reasoning.The refactor in #602 (PR #603) introduced PATO:0102003 (increased) and PATO:0102004 (decreased) as comparative modifier characteristics, on the
has_modifier(RO:0002573) property. Per the uPheno-call agreement of 2026-06-12, this is the direction we want: pre-coordinated qualities defined against the modifier terms rather than against the old increased/decreased grouping classes. I deliberately left the actual sweep out of the upper-level edit and parked it here so #602/#603 stayed focused.Four cases were left half-done in §5 of #602: their
is_alink to the old increased/decreased quality grouping classes (PATO:0002300 / PATO:0002301) was dropped, but nohas_modifieraxiom was added to replace it. They currently only carry their RO:0015007 / RO:0015008intersection_ofaxioms. Those are:Technical note worth flagging, because it changes what this sweep actually does: the RO:0015007 / RO:0015008 axioms are logically independent of
has_modifier some increased/decreased. PATO:0102003 / PATO:0102004 have no equivalence axiom tying them back to RO:0015007 / RO:0015008, so adding the modifier axioms is a genuine semantic addition, not a redundant no-op. That means we cannot assume the sweep is safe by inspection — a reasoner / QC run is required afterwards to confirm nothing goes unsatisfiable.Scope for this issue: enumerate the full set of pre-coordinated value terms, add the
has_modifier some increased/has_modifier some decreaseddefining axioms, and run the reasoner. This should land before curators start composing phenotypes with the new modifier terms, otherwise the compositional and pre-coordinated forms will drift apart.Related