[likhitha-surapaneni]

Test data: NW-52
HGSVC3 data: NW-136
Plants data: NW-148
Accessioned human data: NW-102

Example query:

examples/structural_variation/query.graphql

Testing:

Setup hypsipyle following README.md
Test with new data structure with the example query and check if the response matches example response

Ensembl VDM docs

• https://github.com/Ensembl/ensembl-vdm-docs/blob/main/src/docs/structural_variant.md
• https://github.com/Ensembl/ensembl-vdm-docs/blob/main/src/docs/structural_variant_allele.md

[richysix]

Let me know if any of the comments aren't clear.

Rich

[richysix]

I can't find resolver functions for @STRUCTURAL_VARIANT_ALLELE_TYPE length and copy_number
For copy number it returns null, but for length it causes an error because the schema says length can't be null

query structural_variant_allele_type {
  structural_variant(
    by_id: {
        genome_id: "a7335667-93e7-11ec-a39d-005056b38ce3",
        variant_id: "1:46000:nsv4514092"
    }
  ) {
    name
    alleles {
      name
      copy_number
      length
    }
  }
}

{
  "data": {
    "structural_variant": {
      "name": "nsv4514092",
      "alleles": [
        null,
        null,
        null,
        null,
        null,
        null,
        null,
        null,
        null,
        null
      ]
    }
  },
  "errors": [
    {
      "message": "Cannot return null for non-nullable field StructuralVariantAllele.length.",
      "locations": [
        {
          "line": 12,
          "column": 7
        }
      ],
      "path": [
        "structural_variant",
        "alleles",
        0,
        "length"
      ]
    },
...
}

[likhitha-surapaneni]

Hi @richysix , made a fix to add these resolvers, please let me know if it works as expected

[richysix]

Yep. Looks good

[likhitha-surapaneni]

There's a few other attributes that are in the schema as being NOT NULL, but return null with some of the data files Test data

structural_variant.primary_source: "Cannot return null for non-nullable field StructuralVariant.primary_source.",
structural_variant.slice.region.length: "Cannot return null for non-nullable field Region.length."
structural_variant.slice.region.assembly: "Cannot return null for non-nullable field Region.assembly."

HGSVC3 data

structural_variant.primary_source: "Cannot return null for non-nullable field StructuralVariant.primary_source.",
structural_variant.slice.region.length: "Cannot return null for non-nullable field Region.length."
structural_variant.slice.region.assembly: "Cannot return null for non-nullable field Region.assembly."

structural_variant.alleles.slice.region.length: "Cannot return null for non-nullable field Region.length."
structural_variant.alleles.slice.region.sequence: "Cannot return null for non-nullable field Region.sequence."
structural_variant.alleles.slice.region.assembly: "Cannot return null for non-nullable field Region.assembly."
structural_variant.alleles.slice.region.metadata: "Cannot return null for non-nullable field Region.metadata."

Plants data & Accessioned human data

structural_variant.primary_source.accession_id: "Cannot return null for non-nullable field ExternalReference.accession_id."
structural_variant.primary_source.assignment_method: "Cannot return null for non-nullable field ExternalReference.assignment_method."
structural_variant.primary_source.source: "Cannot return null for non-nullable field ExternalReference.source."

structural_variant.alleles.slice.region.length: "Cannot return null for non-nullable field Region.length."
structural_variant.alleles.slice.region.sequence: "Cannot return null for non-nullable field Region.sequence."
structural_variant.alleles.slice.region.assembly: "Cannot return null for non-nullable field Region.assembly."
structural_variant.alleles.slice.region.metadata: "Cannot return null for non-nullable field Region.metadata."

• Primary source should now work with HGSVC3 and plants data
• For human accessioned, source is not In the header. This seems like a data issue
• For slice.region attributes, out of scope for this PR and will be handled in https://embl.atlassian.net/browse/ENSVAR-6987

[richysix]

Comments are mostly typos and whitespace, but there's one about allele_type for the reference allele in common/file_model/structural_variant.py.

[richysix]

Is this the reference allele?
For SNVs, the reference allele has type biological_region. Should that be the same for structural variants? See my comment in structural_variant.py

[richysix]

For the reference allele, svtype gets set to None, so the if svtype: block doesn't run and the ref allele ends up with a type based on the ALT alleles
For SNVs and small indels, the REF allele type is "biological_region", so I'm assuming that should also happen here.

[likhitha-surapaneni]

Hi @richysix , this should now be fixed. I separated the logic for get_allele_type for structural_variant and structural_variant_allele as it is getting a bit complex. Similar behaviour for short variants

[richysix]

All looks good now.

[likhitha-surapaneni]

Confirm the folder structure with downstream pipeline before merging