Create a Monarch Data Science library

[matentzn]

I am currently thinking of P1 Aim 4 (p1) and general Monarch trajectory, and am thinking of the Monarch Data Science Toolkit:

• Provides a few basic matching services such as alignment with PHENIO: Users upload tables with labels, and obtain SSSOM file with mappings
• Some basic mapping reconciliation (very basic), i.e. get the best set of mappings based on some measure of confidence
• Phenotypic Profile analysis/matching against Monarch KG
• Phenotype similarity based on Monarch KG
• Some query abstractions like monarch.getDiseasesByAnatomicalSite(‘heart’), monarch.getGenesForPhenotype(‘abnormal heart morphology’).
• Enrichment analysis - Flat list of ids (from any domain, disease, anatomy, phenotype, chemical etc) in, enriched report of “whats in the data” out, some kind of profile that collates information from the graph like “parentage” (all terms in set seem to be related to the cardiovascular system), or disease similarity.

I am sure many of you have thought of that, I would think this to be an awesome deliverable for Phenomics first Aim 4. (offline and API modes of course). Is this of interest? is there anyone working on such a library already? I would like to help with organising Workshops involving stakeholders (we have many now we can draw on, from MODs to Vet communities) and making sure the libraries interfaces are designed according to user needs! As this needs considerable FTE to do well (and I would only start it if we do it well), we should decide or reject this now.. Up to you, P1 leadership!