Implement transformation stage for Reactome

[dustine32]

Write inference rules that transform BioPAX data into appropriate GO-CAM structure.

First task will be to add infer_small_molecule_regulators step.

[dustine32]

@tmushayahama For the small_molecule_regulator rule we'll need to update the reactome parser to pass a list of controllers for each reaction. Here's an example rxn with multiple controllers in pathway R-HSA-204174: This reaction ("PDK isoforms phosphorylate lipo-PDH" R-HSA-203946) has multiple entries in it's pybiopax object under controlled_of: Three of them represent the small molecule regulators that are circled at the top of the GO-CAM above. These use Control BioPAX classes. The enabled_by entity of this reaction is in the Catalysis BioPAX class. We should have the parser somehow pass these four controlled_of entries to the GO-CAM data structure used by TransformationStrategy, maybe just simply pass in all controlled_of entries. For these entries, we need to retain the Control and Catalysis type info (since that's used to determine enabled_by vs. is_small_mo_regulator) as well as retain ControlType.

[tmushayahama]

Thanks @dustine32 this is clearer now. Lemme see the biopax. Where is the model R-HSA-204174 on noctua-dev

[dustine32]

Model on noctua-dev: http://noctua-dev.berkeleybop.org/editor/graph/gomodel:R-HSA-204174