Extra causal relations between reaction activities

[dustine32]

This model should only have one causal relation, "directly positively regulates" (RO:0002629), between these two reaction activities:

Reaction 1: R-HSA-1504186
Reaction 2: R-HSA-201677
Pathway: R-HSA-4641262

Relation "provides input for" (RO:0002413) should not have been added. Current hypothesis is:

1. Relation "directly positively regulates" is added in the inferRegulatesViaOutputEnables rule.
2. Pre-existing relation "causally upstream of" (RO:0002411) is removed or should have been removed (or is added back again later) because it is a more generic relation than "directly positively regulates."
3. Rule inferProvidesInput is run, which detects this reaction pair because "causally upstream of" is somehow still connected between them. The rule matches output "WNT:FZD:LRP5/6:DVL:AXIN:GSK3B" from rxn 1 and input from rxn 2 and replaces the "causally upstream of" relation with "provides input for".

Find out why "causally upstream of" is not removed in step 2.

[dustine32]

@ukemi @deustp01 I just realized this situation may be allowed according to this comment from Ben:

pathways2GO/exchange/src/main/java/org/geneontology/gocam/exchange/GoCAM.java

Lines 1502 to 1503 in 6ffeb27

	//don't re-initialize the query model yet. If we remove the causally upstream of relation, the next rule can't fire
	//and it is possible for a reaction to both catalyze and provide input for a another reaction.

So, is it indeed fine for there to be two causal relations between reactions here to "to both catalyze and provide input for a another reaction"?