Add UDP models

[julesjacobsen]

@mellybelly to fill in the details of the data source.

[damiansm]

Whats the idea here? To add solved gene to phenotype associations from UDP or UDN? Is there a decent number now and are they really sharable with the world?

[mellybelly]

The idea is to amass a number of N-of-1 public cases, such as those shared by the UDP to Monarch, @kshefchek can provide, and new ones coming from all of UDN, such as their patient pages: https://undiagnosed.hms.harvard.edu/participants/participant-001/ We will be exporting these. Its a great use case for phenopackets @pnrobinson The idea is to allow user to include/exclude n-of-1 diseases orphanet wants these too, and monarch already has public pages for these. no reason not to match these in the more widely use Exomiser rather than rely upon MME, for those that have been made public. We can also get the MyGene2 ones as well (though no API there AFAIK)

[damiansm]

Sounds cool. Guess we need to think about user options/defaults for these n-of-1 cases. Are they solved, possible candidates or a mixture? Guessing the latter. Of course OMIM already has some n-of-1 weak associations anyway so we should probably not be too scared with using this data as long as properly flagged by source.

[kshefchek]

The variant information is sometimes lacking, we attempt to resolve snps and indels to dbSNP, or make informative HGVS labels, but YMMV.

Here are two example queries: https://solr.monarchinitiative.org/solr/golr/select/?q=*:*&wt=json&fq=subject_category:case&fq=object_category:variant&fq=is_defined_by:%22https://data.monarchinitiative.org/ttl/udp.ttl%22&rows=20&indent=on&fl=subject,object,object_label

https://solr.monarchinitiative.org/solr/golr/select/?q=*:*&wt=json&fq=subject_category:case&fq=object_category:phenotype&rows=20&indent=on&fl=subject,object

[mellybelly]

A mixture. agreed- we just need flags - both for "n-of-one" and for source