just_dna_format.consequences¶
just_dna_format.consequences ¶
The per-transcript consequence table: what a variant does to each transcript near it (0.8, RM236).
consequences.csv is a derived-fact sidecar filled by just-dna-enricher's opentargets enrich pass.
One row per (variant_key, transcript), which is the grain RM236 chose over the two cheaper shapes:
- one MANE value per variant is lossy, and a column shipped under a major cannot be retyped into a table later;
- the most severe consequence across transcripts is a ranking policy, which is an interpretation and stays the consumer's.
RM23's computational predictor scores have the same blocker (a score is per transcript), and inherit this grain when they are built; one answer to "which transcript" rather than two.
The transcript set is the source's choice, and the description says so. Open Targets 26.09 lists
one transcript per nearby gene, every one flagged Ensembl-canonical (probes/LANES_0_8.md § 1.6), not
every transcript of every gene. canonical records the flag rather than the table implying it.
The consequence is stored as Sequence Ontology accessions, verbatim (SO_0001583), because that
is what the source publishes. A term label (missense_variant) would be a mapping this format would
own, and the name consequence stays free for it.
ConsequenceRow ¶
Bases: BaseModel
One (variant, transcript) pair: the predicted consequence and its impact class.