search

intermine / pombemine

0 stars 1 forks source link

modelling MONDO associations to human (include the entire dataset, not our curated associations) #53

Open ValWood opened 2 years ago

ValWood commented 2 years ago

PomBase have made curated associations to MONDO terms. It would be useful to also load instead all MONDO associations to human proteins. We could use this query to cross-check our associations by joining on the orthologs using a) the manually curated ortholog associations and b) the PANTHER orthologs once available

ValWood commented 2 years ago

i.e model the MONDO associations exactly the same as the OMIM ones:

Screenshot 2022-05-23 at 20 35 11

just as an additional "Dataset."

There is no need to load the PomBAse curated associations if this is problematic. I will just upload our list list and use the Mine to find any inconsistencies. We planned to migrate to an automated pipeline to maintain the MONDO associations soon going forwards now we have done the DO->MONDO migration. So now I understand more how the Mine works makes sense to aliign with the long term plan and just load the disease associations from the primary source.

If the manual cuarration of these can

danielabutano commented 2 years ago

@ValWood why not use this query? image

ValWood commented 2 years ago

There is a long an complicated history about the disease associations. we have been curating them since before disease ontologies. Then for over a decade or so we used DO. In 2020 we switched to MONDO. In the early stages we could not find a good mapping of the diese ontology terms to human genes. However Monarch now provides this connection in an accessible way. I think we would be better to migrate from our curated list to updating automatically from Monarch. So it would make sense that instead of using PomBAse gene -> MONDO (via our curated ortholog table) just loaded MONDO -human connections into PombeMine so that we could query them alongside OMIM since they are the same data type from different sources (OMIM and Monarch). This would mean that as new pombe-human ortholog are identified the MONDO data would already be in the mine

I can only access the OMIM dataset here:

Screenshot 2022-05-27 at 16 24 18

but ideally it would be both MONDO (via MONARCH and linking to human orthologs), so that I can evaluate them both together (instead of our curated subset loaded in a different place).

I can see MONDO in the list here, but I am unable to add to the same table as the OMIM data.

If the MONDO -human disease association can be loaded directly from MONARCH, we would stip exporting our sub optimal curated dataset.

Sorry I don't think I am explaining this very well!

ValWood commented 2 years ago

OK it seems in the timescalse we won't be able to do this because it would require a Monarch loader and this does not seem to exist. I had assumed that the Alliance used Monarch/MOndo but they are still using DO.