OncoRef 1.27.0: simpler decisions and adult Disposable evidence
701 CTA / 18,062 Unsafe / 453 Other / 647 Unassessed representative proteins. All 701 previous CTAs and all 18,062 previous Unsafe assignments are preserved.
Simple rules, funnel and four target reviews · Full source report · Rules and API · PDF logic slides
One decision, separate evidence fields
The main decision is CTA, Unsafe, Other or Unassessed. Other now records only its unmet identity requirement: 448 Nonreproductive / 5 Diffuse; no current protein fails only cancer presence. The former Low, Weak, Normal, Tumor-supported and Disposable labels mixed abundance, confidence and anatomy.
cta_selection.profiles() exposes those evidence dimensions independently:
separate native TPM/nTPM/pTPM/CP10K summaries, replicated tumor RNA, normal protein,
reproductive presence, developmental measurements and adult Disposable expression.
The top-level result and identity fields remain available through results().
Update old subtype filters explicitly; the new subtype names describe reasons.
Disposable is an adult tissue disposition
An adult compartment's permission never depends on embryonic/fetal/placental expression. The adult Disposable-only expression attribute now recognizes TCHHL1, LCE3B and LCE3C among Other proteins; developmental measurements remain separate, including missing data. The existing adult anatomy mapping and safety ceilings are preserved. Whole skin remains Manageable; mapped follicles and cutaneous melanocytes have conditional Disposable allowances.
One additional normal-evidence hold
OR10A4 moves from Other to Unassessed. Original human olfactory-mucosa data show
RNA across 26 adult donors, and a separate study detects normal nasal-fluid RNA.
These native qPCR/copy measurements are not TPM or protein. They prompt review of
a poorly covered normal compartment, rather than a claimed clinical toxicity.
The two source spreadsheets are pinned through the shared catalog / reference_data
API at version 2014-05-06. Measurements, citations and limitations.
DBX1, OR2B2 and KRTAP20-2 retain strong tumor evidence but lack qualifying reproductive identity. Their atlas normal peaks and early-developmental values are in the report. High tumor expression alone does not make a protein a CTA.
Downstream adoption
from oncoref import cta_selection, catalog
results = cta_selection.results()
profiles = cta_selection.profiles()
roles = cta_selection.sequence_dispositions()
normal_source = catalog.ensure("olfactory_verbeurgt_cq", version="2014-05-06")
Retain policy/reference identity, raw evidence and sequence QC. OR10A4 now requires review for shared-sequence allowance. Historical FASTA/k-mer bundles remain separately versioned; see the unresolved sequence-bundle and translation-QC contracts in #641 and #657. Human policy does not transfer to dog/mouse; their profile measurements stay missing.