EVA-2352 — ClinVar investigations

[tskir]

For the report, it's easier to view the updated version rather than read a diff: https://github.com/EBIvariation/eva-opentargets/tree/eva-2352-clinvar-investigations/clinvar-variant-types.

• Closes Investigate TraitSet #199. New diagram: “Trait representation”.
• Closes Investigate issues with mode of inheritance values in ClinVar #157. Refactored diagram and added table: “Mode of inheritance”.
  • Now divided into more sensible four categories: missing, germline, somatic, germline & somatic.
  • Multiple values are output into a separate table.
• Closes Improve allele origin #163. Refactored diagram and added table: “Allele origin”. Similarly to mode of inheritance, it's divided into four categories, and multiple values are output into a separate table.
• Closes Discrepancy between origin and mode of inheritance #186. Closes Investigate relationship between allele origins and mode of inheritance #190. New diagram and table: “Mapping between mode of inheritance and allele origin”.
  • Displays relationship between the four high level categories of mode of inheritance and allele origin.
  • Discordant records are output into a separate table.
• Closes Refactor the clinvar-variant-types script to use all clinvar_xml_utils functionality #170. Very thorough refactor of the ClinVar investigations script:
  • Now consistently using functionality provided by clinvar_xml_utils.
  • Added new classes for Sankey diagrams and tables. This makes adding new diagrams and tables in the future significantly easier.

This PR only concerns the updated script and the report. Separate follow-up issues will be created to discuss the findings with the Open Targets team and plan future work, if necessary.

[tcezard]

This is a very nice addition to the parsing libraries.
It illustrates really well the different types of submission that clinvar contains

[tcezard]

It might be worth adding a diagram of the different trait types as well.

[tskir]

You're right, that's an important metric. Thank you for this suggestion, I've now implemented it, and the diagram really looks much more informative. Here's an updated section in the report: https://github.com/EBIvariation/eva-opentargets/tree/eva-2352-clinvar-investigations/clinvar-variant-types#trait-representation