For YeastPathways, use stepDirection instead of conversionDirection

[dustine32]

Currently, the pathways2go code will only use the conversionDirection field on BiochemicalReaction to determine how to arrange the left and right molecules in an activity. In YeastPathways BioPAX, the conversionDirection field is only used when a reaction is REVERSIBLE - direction information is primarily in the BiochemicalPathwayStep stepDirection field.

For YeastPathways, we should use stepDirection to give us the direction of the reaction in the pathway, and ignore conversionDirection. Some example models/reactions where both are present and contradict each other:

1. PENTOSE-P-PWY
   • StepDirection LEFT_TO_RIGHT not equal to ConversionDirection REVERSIBLE in TRANSALDOL-RXN
2. FOLSYN-PWY-1
   • StepDirection LEFT_TO_RIGHT not equal to ConversionDirection REVERSIBLE in FORMATETHFLIG-RXN
3. NONOXIPENT-PWY
   • StepDirection RIGHT_TO_LEFT not equal to ConversionDirection REVERSIBLE in RIB5PISOM-RXN

Tagging @thomaspd

[dustine32]

@suzialeksander @thomaspd Here's the full list of pathways having a reaction step with stepDirection = RIGHT-TO-LEFT: https://docs.google.com/spreadsheets/d/1bbDSJ5uNhgWhCg4y6llfGazKbMGILsjz2J3hsuU1B5Y/edit#gid=0

These are currently loaded into noctua-dev and the stepDirection fix is ready for testing. Enjoy!

[dustine32]

@thomaspd @suzialeksander Just checking if any changes need to be made for this stepDirection code. If not, I can merge the PR and close this ticket. Noctua-dev still has the test models with this change loaded (example: NONOXIPENT-PWY).

[suzialeksander]

Hey @dustine32, this sounds correct. No objections here!

[dustine32]

Wonderful, thank you @suzialeksander !