Genome-wide normal-expression categories

Policy oncoref.normal-expression-categories.2026-10-21.2; 20,097 canonical human protein-coding genes, pinned HPA v23. This universe includes genes outside published CTA nominations.

Category Other subgroup Genes
Normal — 18,063
Other Disposable 29
Other Silent 372
Reproductive — 1,023
Unassessed — 610

Definitions

  • Reproductive: positive reproductive RNA >0 or IHC, complete routine RNA, and all approved absolute RNA/selectivity/IHC gates. Publication nomination and manual CTA review are independent: this is broader than the retained curated CTA set.
  • Normal: measured expression fails an absolute ceiling, the positive-IHC gate, or (when reproductive-positive) the pairwise selectivity gate. It means normal-expression counterevidence under this policy, not demonstrated clinical toxicity.
  • Other: no positive reproductive evidence, complete routine RNA, no disqualifying positive IHC, and all approved absolute category/IHC ceilings pass. No reproductive ratio is required. Disposable has a measured positive in Disposable tissue; Silent does not. Both mean below-threshold expression, not literal absence from every normal tissue.
  • Unassessed: no decisive normal-expression counterexample, but missing routine RNA, unresolved anatomy, unusable/reviewed-unreliable evidence, or an undefined required reproductive ratio prevents classification into Reproductive/Other. It is a separate evidence status and never part of Silent.

Missing/partial IHC retains baseline RNA limits and is explicitly recorded; this does not establish protein absence. The 50 RNA/80 IHC routine contexts do not cover all developmental stages or anatomical compartments. Numerical limits apply only to human HPA v23. Exact category/IHC grid.

Preparing for subsequence comparisons

The three main categories are disjoint research screening groups. Sequence membership is a separate question: a reproductive gene can share a subsequence with a normal paralog, and normal RNA below a screening limit is still observed expression.

normal_expression_observed preserves any valid RNA >0 or positive IHC in Manageable/Protected/Vital tissue, even below a cutoff. The conservative normal-reference export includes those genes, all Normal genes, and Unassessed genes. It can overlap Reproductive and Other. A later subsequence comparison must examine exact protein/transcript occurrences and their reference versions across both sets; this report does not establish sequence specificity. Disposable expression remains separately visible.

Downloads

The CTA recalibration report applies publication, family and nomination reviews in addition to the Reproductive expression screen. The current doubled absolute ceilings preserve every previously passing Reproductive expression decision. retained_recalibration_cta records that pre-external-review stage; use the current clinical curation decisions for the final retained set and external safety holds.

Reproduce with python scripts/normal_expression_categories_report.py --out <empty-directory> --workers 4.