Updates for GnomAD variant annotation

[DSuveges]

As part of the new variant page effort a series of changes should be implemented in the GnomAD based variant annotation. This list is expected to be extended until we run out of updates, then upon prioritizing the task, all should be implemented.

• [x] Moving various identifiers into a cross-references object [1].
• [ ] Related: provide corss reference to protvar when the predicted consequence of variants make sense [4].
• [ ] Moving amino-acid change for missense variants out of VEP object and move it into the root [2].
• [ ] Adding alphafold based predicted impact of amino-acid change to the insilico predictor object [3].

[1] Cross references: Is a list object with struct elements. id provides the identifier in the cross referenced database, source is defines the external dataset. The following example coming form the target object:

      ├───dbXrefs: array 
      │   ├───element: struct 
      │   │   ├───id : string
      │   │   ├───source : string

[2] Currently the amino-acid change is inside the transcript VEP object. This needs to be found and extract to root level.

[3] We have a predicted mutation effect for evey amino acid cange in the full human proteom based on alphafold predicted structures. Finding the annoation is not trivial. Needs to be joined by uniprot id, position, reference amino acid, alternate amino acid.

[4] Protvar provides predicted consequences and annotations for variants causing amino acid changes in proteins. We have to collect those consequence terms, where protein chain alteration happens and construct link to protvar. An URL can look like this: https://www.ebi.ac.uk/ProtVar/query?search=10-43118436-A-C

[DSuveges]

• Remove |-- gnomadVariantId: string (nullable = false) from existing variant annotation.
• Remove GRCh37 mappings as well. We no longer use or show.

[DSuveges]

I'm closing this issue, as we are departing from GnomAD as a source of variant annotation.