Add pathogenicity predictions and allele to frequency

[costero-e]

This is the work that @dglemos has been working on and already deployed in the genomicVariations feature branch. This is ready to be merged into the develop branch IMO as the work seems finished.
Can you please review, @mbaudis and @redmitry ?

[redmitry]

Since it was merged, I have updated my java model... a month ago.

https://gitlab.bsc.es/inb/ga4gh/beacon-v2-api/-/blob/master/beacon-data-model/src/main/java/es/bsc/inb/ga4gh/beacon/model/v200/variations/PopulationFrequency.java#L41
https://gitlab.bsc.es/inb/ga4gh/beacon-v2-api/-/blob/master/beacon-data-model/src/main/java/es/bsc/inb/ga4gh/beacon/model/v200/variations/VariantLevelData.java#L43

D.

[jrambla]

I have some questions about the suggested changes. I know the PR has been waiting for some months.

1. Why do we need to add the allele if, actually, the section is inside a given genomicVariation?
2. The pathogenicityPrediction should be linked to an specific phenotypic effect, is not generic to any condition.
3. The "phred_score:0.381" example doesn't seem compliant with the ontologyTerm, is it?

[costero-e]

@dglemos Can you check @jrambla questions?

[dglemos]

Hi @jrambla,

For the following query: assemblyId=GRCh38;referenceName=9;start=22125503;referenceBases=G;alternateBases=C
The Ensembl Beacon returns a match: rs1333049 -> this variant has two alt alleles A and C. As some of our data is not attached to a specific allele (only attached to ID rs1333049) our beacon returns both alleles in the response.
This is how the alleles are returned in the response:

"variation": {
              "referenceBases": "G",
              "alternateBases": "A,C",
              "variantType": "SNP",
              "location": {
                "interval": {
                  "type": "SequenceInterval",
                  "end": {
                    "value": 22125504,
                    "type": "Number"
                  },
                  "start": {
                    "type": "Number",
                    "value": 22125503
                  }
                },
                "type": "SequenceLocation",
                "sequence_id": "9"
              }
            }

1. We prefer to attach the specific alt allele associated with the frequency in the response. This way we make sure the user knows which one has the frequency.
2. The same applies here. We don't have pathogenic predictions directly linked to phenotypic effects.
3. I'll review this one.