Add warning for when a coordinate occurs outside of 100 base pairs of the start/end of a transcript

[katiestahl]

Feature description

We should have a warning for when a coordinate occurs outside 100 base pairs of the start or end of a transcript.

Example:
a user could call genomic_to_tx_segment with

genomic_ac="NC_000001.11",
seg_start_genomic=99999999999999999999999,
gene="TPM3",
get_nearest_transcript_junction=True

and it would return results. (below output is from calling api/utilities/get_exon?chromosome=NC_000001.11&start=99999999999999999999999999999&gene=TPM3 from the fusion-builder api, which calls genomic_to_tx_segment directly.)

{
  "coordinates_data": {
    "gene": "TPM3",
    "genomic_ac": "NC_000001.11",
    "tx_ac": "NM_152263.4",
    "seg_start": {
      "exon_ord": 9,
      "offset": -1e+29,
      "genomic_location": {
        "type": "SequenceLocation",
        "sequenceReference": {
          "type": "SequenceReference",
          "refgetAccession": "SQ.Ya6Rs7DHhDeg7YaOSg1EoNi3U_nQ9SvO"
        },
        "end": 1e+29
      }
    },
    "errors": [],
    "service_meta": {
      "name": "cool_seq_tool",
      "version": "0.7.1",
      "response_datetime": "2024-09-26T17:13:08.423570Z",
      "url": "https://github.com/GenomicMedLab/cool-seq-tool"
    }
  }
}

Use case

As a user, I might accidentally enter an incorrect coordinate that gets resolved to the nearest transcript junction and get some odd results that may be confusing. I would benefit from a message warning me that my entered data or the returned data could be unexpected.

Proposed solution

Add a warning when the coordinate occurs so many base pairs away from the start/end of a transcript

Alternatives considered

No response

Implementation details

When genomic_to_tx_segment is called with get_nearest_transcript_junction=True, return an error if the coordinates occur over 100 base pairs away from the transcript start / end

Potential Impact

No response

Additional context

Example of response from calling genomic_to_tx_segment with erroneous coordinates

you can see that the entered coordinate and corresponding offset will scale infinitely

Contribution

Yes, I can create a PR for this feature.