new: "Understanding signaling and metabolic paths using semantified and harmonized information about biological interactions"

I joined the BiGCaT research team because of their open science work on WikiPathways (Scholia summary). To me, WikiPathways is an essential tool to move forward. It's different from Reactome and KEGG and I think for good reasons (read the articles to get some idea). I also believe collaboration with both databases is very useful. I have seen how open science cheminformatics boosted the field in general, and do not see how that would not work here too. With Reactome we have a working collaboration (doi:10.1371/journal.pcbi.1004941) but this is also why I think ELIXIR is important, which encourages such collaborations much more widely.

Ryan Miller just published work on the interactions in WikiPathways: Understanding signaling and metabolic paths using semantified and harmonized information about biological interactions (doi:10.1371/journal.pone.0263057). We already had some good understanding about the harmonization of data nodes (genes, proteins, metabolites) in the pathways. Here, BridgeDb is an important component and the databases identifiers help software understand the biological meaning of a data node in a pathway drawing.

However, for interactions the situation is more complex. First, it relies on the database identifiers of the nodes. For example, are these consistent with each other? And what if we consider the meaning of the interaction? Is it ligand binding? A protein-protein interaction? Do the database identifiers still make sense then?

Another complications is actually that WikiPathways is more liberal with what biology authors describe. That leads to a higher amount of variation of drawing styles. Some of these are formalized in standards like MIM and SBGN (see above example for inhibition). Can this be harmonized? I guess it is clear how important understanding these aspects are if you use the information in network analyses, something we regularly do. 

Ryan developed a framework to keep track of these things and along the way applied it to the WikiPathways release of June 2019. I'm quite happy this work is now published: it gives us something (a scalable and semantic approach) we can build upon.

Happy reading!