No results when specify transcript, but works for "raw" (VV new version)

[lcoutos]

Hi,
I faced an issue with the Formatter with the new version of VariantValidator (I had no problem with the same variant with the previous version). For many variants, I had no results for Formatter for old refseq transcripts (from cbioportal data), but I found the transcript with gene2transcripts, and the variant was correctly validated with VariantValidator on the same transcript. After few tests, I realized that my transcript was correctly managed if I set specify_transcripts = "raw" (no differences between GRCh37 and GRCh38). For exemple:

Recent transcript, works perfectly

specify_transcripts = "NM_000546.6"
variant = {"genesymbol": "TP53", "hgvsg": "NC_000017.10:g.7578449C>T", "assembly": "GRCh37"}
formatter_result = VariantFormatter.simpleVariantFormatter.format(batch_input = variant["hgvsg"], genome_build = variant["assembly"], transcript_model = "refseq", specify_transcripts = specify_transcripts)  

Output:

{
  "NC_000017.10:g.7578449C>T": {
    "errors": [],
    "flag": null,
    "NC_000017.10:g.7578449C>T": {
      "p_vcf": "17:7578449:C:T",
      "g_hgvs": "NC_000017.10:g.7578449C>T",
      "selected_build": "GRCh37",
      "genomic_variant_error": null,
      "hgvs_t_and_p": {
        "NM_000546.6": {
          "t_hgvs": "NM_000546.6:c.481G>A",
          "p_hgvs_tlc": "NP_000537.3:p.(Ala161Thr)",
          "p_hgvs_slc": "NP_000537.3:p.(A161T)",
          "select_status": {
            "refseq_select": true,
            "mane_select": true
          },
          "gene_info": {
            "symbol": "TP53",
            "hgnc_id": "HGNC:11998"
          },
          "transcript_version_warning": null,
          "gapped_alignment_warning": null,
          "gap_statement": null,
          "transcript_variant_error": null
        }
      }
    }
  },
  "metadata": {
    "variantvalidator_version": "3.0.2.dev57+gf344e4e.d20250402",
    "variantvalidator_hgvs_version": "2.2.1.dev7+g59392c5",
    "vvta_version": "vvta_2025_02",
    "vvseqrepo_db": "VV_SR_2025_02/master",
    "vvdb_version": "vvdb_2025_3",
    "variantformatter_version": "3.0.1.dev13+ga2cbf30"
  }
}

Old transcript, no result

specify_transcripts = "NM_001126112.2"
variant = {"genesymbol": "TP53", "hgvsg": "NC_000017.10:g.7578449C>T", "assembly": "GRCh37"}
formatter_result = VariantFormatter.simpleVariantFormatter.format(batch_input = variant["hgvsg"], genome_build = variant["assembly"], transcript_model = "refseq", specify_transcripts = specify_transcripts)  

Output:

{
  "NC_000017.10:g.7578449C>T": {
    "errors": [],
    "flag": null,
    "NC_000017.10:g.7578449C>T": {
      "p_vcf": "17:7578449:C:T",
      "g_hgvs": "NC_000017.10:g.7578449C>T",
      "selected_build": "GRCh37",
      "genomic_variant_error": null,
      "hgvs_t_and_p": {
        "intergenic": {
          "alt_genomic_loci": null
        }
      }
    }
  },
  "metadata": {
    "variantvalidator_version": "3.0.2.dev57+gf344e4e.d20250402",
    "variantvalidator_hgvs_version": "2.2.1.dev7+g59392c5",
    "vvta_version": "vvta_2025_02",
    "vvseqrepo_db": "VV_SR_2025_02/master",
    "vvdb_version": "vvdb_2025_3",
    "variantformatter_version": "3.0.1.dev13+ga2cbf30"
  }
}

specify_transcripts to "raw', my old transcript NM_001126112.2 is in result, and correctly managed

specify_transcripts = "raw"
variant = {"genesymbol": "TP53", "hgvsg": "NC_000017.10:g.7578449C>T", "assembly": "GRCh37"}
formatter_result = VariantFormatter.simpleVariantFormatter.format(batch_input = variant["hgvsg"], genome_build = variant["assembly"], transcript_model = "refseq", specify_transcripts = specify_transcripts)  

Output (only a short extract):

"NM_001126116.2": {
          "t_hgvs": "NM_001126116.2:c.85G>A",
          "p_hgvs_tlc": "NP_001119588.1:p.(Ala29Thr)",
          "p_hgvs_slc": "NP_001119588.1:p.(A29T)",
          "select_status": {},
          "gene_info": {
            "symbol": "TP53",
            "hgnc_id": "HGNC:11998"
          },
          "transcript_version_warning": null,
          "gapped_alignment_warning": null,
          "gap_statement": null,
          "transcript_variant_error": null
        }

I can still run with specify_transcripts to "raw" when necessary, so it's not a huge problem and there is no emergency.
Thanks!