Different number of Reactome pathways in processed/original/intermediate BioPAX files

24 views
Skip to first unread message

Julian Kreis

unread,
Feb 24, 2021, 1:21:20 AM2/24/21
to pathway-commons-help
Hello,

Thanks a lot for your great efforts and this great web service! I checked your final Biopax file (PathwayCommons12.reactome.BIOPAX.owl.gz), and the different intermediate files in the original-intermediate-data/reactome folder (*original/*/normalized/*cleaned). While the final file has 2,203 pathway names annotated in the PATHWAY_NAMES columns, the original and intermediate files only have 1,503. However, the Data Sources website states, that there are 2,272 pathways. The reactome v75 release (Homo_sapiens.owl) has 1,584 pathway names annotated, but reactome states that there are 2,477. Why are the number of pathways different? Additionally, could you point me to a description of your processing procedure?

Best,
Julian

Igo -

unread,
Feb 24, 2021, 9:35:48 AM2/24/21
to pathway-co...@googlegroups.com, Gary Bader
Hi Julian,

At that time, we processed the Reactome v69, 28-May-2019, only 'Homo_sapiens.owl' (not v75). Pathway Commons project is not maintained anymore (funding, resources).

Where did you see the PATHWAY_NAMES column (there is no such property in BioPAX model)? Also, a Pathway can have multiple, even duplicate names.
Anyway, to count the number of pathways in a BioPAX model, we simply count the number of owl:Pathway objects - each has an URI (rdf:ID - partial URI) and displayName (not always present, nor unique, I think).
There are also non-human (some human virus) pathways in the Reactome Home sapiens model. Thus 2,272. We used a data driven approach.

The numbers reported in this file and on the PC2 page, were calculated from the final BioPAX model (PathwayCommons12.All.BIOPAX.owl.gz). We simply count all the owl:Pathway objects in a file/model, I think.
(The number of other BioPAX entity or utility classes may differ from those in the original/intermediate files due to normalization and merging the models into one.

You can also check it as follows (get paxtools.jar that we used here; see also https://biopax.github.io/Paxtools/, run as "java -jar paxtools.jar" for help):
java -Xmx8g -jar paxtools.jar summarize PathwayCommons12.reactome.BIOPAX.owl.gz out.txt 
less out.txt | grep Pathway (it shows 2272 for all those Reactome files, incl. the intermediate data)

For more info, please visit Github and other links below:
Best regards,
Igor R.


--
You received this message because you are subscribed to the Google Groups "pathway-commons-help" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pathway-commons-...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/pathway-commons-help/1f5abf3a-55fd-47b9-92ea-11f1672e3b17n%40googlegroups.com.
Reply all
Reply to author
Forward
0 new messages