Finer control over what goes into BioPAX bp:Pathway 'comments'

[jvwong]

Part question, part update: Is there any way to control what gets into a bp:Pathway comment property, in particular, from Biofactoid?

1. Hiding comments

In the beta PC instance instance, most of the Biofactoid BioPAX bp:Pathways have comments like REPLACED http://pathwaycommons.org/pc13/Pathway_.... See example LD Motif Recognition by Talin: Structure of the Talin-DLC1 Complex.

2. Customizing comments

• There's at least two types of comments that would be nice to have show up:
  • a. Reference to the paper that was tied to the Biofactoid document (e.g. for above: Zacharchenko et al. LD Motif Recognition by Talin: Structure of the Talin-DLC1 Complex., Structure (2016))
  • b. We have started to capture free text 'context' in Biofactoid, and it would be useful to put those somewhere (see https://github.com/PathwayCommons/factoid/issues/1013)

[metincansiper]

1. Hiding comments

Looks like REPLACED http://pathwaycommons.org/pc13/Pathway_... is not coming from biopax converter service, I guess it must be added by the cpath2. I searched the word "REPLACED" in all the biopax documents that is exported from biofactoid and see that none of them contains it. Also, having pc13 in the string made me suspect that it must be coming from cpath2.

2. a.

Currently, we set the displayName property of the Pathway as the title of the citation. For this case it is LD Motif Recognition by Talin: Structure of the Talin-DLC1 Complex. Do you want to have both this displayName and the comment as you mentioned (including the title inside but broader).

[jvwong]

1. Hiding comments

Looks like REPLACED http://pathwaycommons.org/pc13/Pathway_... is not coming from biopax converter service, I guess it must be added by the cpath2. I searched the word "REPLACED" in all the biopax documents that is exported from biofactoid and see that none of them contains it. Also, having pc13 in the string made me suspect that it must be coming from cpath2.

OK. If it's not the concern of this code, let's leave it for now.

2. a.

Currently, we set the displayName property of the Pathway as the title of the citation. For this case it is LD Motif Recognition by Talin: Structure of the Talin-DLC1 Complex. Do you want to have both this displayName and the comment as you mentioned (including the title inside but broader).

Now that I think about this, items like the title, reference to the paper should probably be the domain of the app-ui system to display / format information.

I still think a valid update is adding the Biofactoid 'context' strings to the bp:comments for the pathway (unless there's a better place for this).

[IgorRodchenkov]

Hi everyone. I think those "REPLACED" come from the cpath2 "merger" or "normalizer" process. Yeah, some of these are not necessary (when the RDFId is in fact not even updated), sort of a bug. As to what must go into comments, I think, usually there're some comments in the original data; I would not add there publication title or ID as these should go with publication xrefs (and can be later extracted via pubmed web services or alike - like we do to display the pub details on the PC "Datasources" web page).

[IgorRodchenkov]

I think this procedure call somehow misfires for the Pathways... https://github.com/PathwayCommons/cpath2/blob/master/src/main/java/cpath/service/Merger.java#L98 (perhaps the xmlBase did not match at the time the tool was run.. or some other hidden bug)