Deal with protein modifications

[tythomson]

Many reactions in reactome involve phosphorylation, or other post translational modifications. e.g. http://www.reactome.org/PathwayBrowser/#/R-HSA-163685&SEL=R-HSA-163773&PATH=R-HSA-1430728

Ideally we could capture the specific modification and site information in BEL using pmod() - for the example above this would be p(HGNC:PFKFB1, pmod(P, S, 33)). If this is challenging, an initial step towards this might be to capture more generally the modification without site information, e.g., p(HGNC:PFKFB1, pmod(P)).

Currently all post translational modifications are lost. Many proteins only undergo a single modification within the reactome network, so capturing the information without the site would be extremely helpful for a large number of reactions.

[tythomson]

Here's an example of a post translational modification that may be hard to represent in BEL. It's a nice edge case to consider, because if it can't be accurately represented we might want to throw it out entirely rather than have some weird representation.
http://www.reactome.org/PathwayBrowser/#/R-HSA-163685&SEL=R-HSA-5228508