CTA screening: retained, variable and excluded
Updated 23 October 2026. 740 retained, 1 Variable (excluded now) and 1799 other exclusions among 2,540 mapped candidates. 18 former point-screen passes are removed; no new candidates are admitted.
Retained is a research-policy pass, not clinical clearance. Only CTAG2 currently has imported individual-donor RNA evidence. All other donor profiles remain Not assessed.
- Find a gene — filter by category and open its evidence.
- Simple rules · CTAG2 and Variable · Named targets.
- All gene results · Retained set · Variable set · Changes.
- Publication intake and unmapped nominations · Previous 758-gene pilot.
The rules, in order
| Rule | Action |
|---|---|
| No reproductive expression, unresolved identity/coverage, family exclusion, or a separate safety hold | Do not select. Preserve the reason. |
| Positive IHC, including Low, in a Vital, Protected or Manageable compartment | Exclude. A positive reproductive control cannot override it. |
| Missing/uncertain normal IHC, incomplete local staining, or no reliable positive reproductive IHC | Use baseline RNA limits. Missing staining is never a negative result. |
| Complete reliable negative local IHC and reliable positive reproductive staining | Allow the existing confidence-dependent RNA relaxation. |
| Any screened tissue fails its RNA ceiling, selectivity, or reproductive-higher rule | Exclude. |
| A reviewed individual donor fails those same RNA rules in a Vital tissue | Exclude even when the tissue average passes. Classify the qualifying zero/high, IHC-negative pattern as Variable. |
Reproductive positivity is checked at the mapped cell compartment. Disposable expression is ignored. The positive control is currently established within the curated HPA gene profile: the export does not prove identical antibody, batch or specimen across every tissue. It does not calibrate antibody sensitivity or a proportional RNA-to-protein relationship.
The numerical limits
| Tissue category | Absolute nTPM ceilings | Selectivity minima |
|---|---|---|
| Vital | 2 / 4 / 6 / 12 | 80 / 75 / 70 / 65% |
| Protected | 6 / 8 / 12 / 24 | 75 / 70 / 65 / 60% |
| Manageable | 10 / 20 / 30 / 60 | 70 / 65 / 60 / 55% |
Order: baseline → Approved-negative → Supported-negative → Enhanced-negative. Every relaxation requires the reproductive positive control above. Selectivity is tightened by five percentage points throughout, as tested against the named panel; absolute ceilings are unchanged. These are research settings, not validated safety limits.
For each tissue, let R be the highest reproductive nTPM, N its normal nTPM, C its absolute ceiling and s its selectivity minimum. Require N ≤ C, N ≤ R × (1−s)/s + 0.1, and N < R. The small continuous allowance stays fixed at 0.1 nTPM. Reviewed donor values replace N in exactly these same comparisons. There is no separate 10 nTPM donor cutoff.
Variable: excluded now, reconsiderable later
Membership requires:
- At least one reported-zero RNA donor and a different RNA donor failing the applicable limits, within the same gene, Vital tissue, cohort, source version and units.
- Complete reliable negative IHC in the affected tissue's routinely measured compartments; no competing normal-protein exclusion or separate safety hold.
- Measured reproductive expression and usable candidate identity/coverage.
Variable never means selected. It records a reason to revisit the exclusion. No measurement becomes zero because it is missing, and zeros/high values from different cohorts are not combined to manufacture membership. Negative IHC in other volunteers does not clear an RNA-positive donor. A patient with an RNA-zero measurement is not automatically eligible.
CTAG2 / LAGE-1
- Pinned v23 heart RNA donors: 13.5, 0, 0 and 7.1 nTPM. The consensus heart estimate is 5.1 nTPM; reproductive maximum is testis 10.1 nTPM.
- Heart has Enhanced-negative cardiomyocyte IHC and testis has positive reproductive staining. The applicable absolute ceiling is 12 nTPM and the selectivity-derived ceiling is 5.538 nTPM. Thus both nonzero donors fail, while the mean passes.
- Three separately sampled heart IHC donors (women aged 52/62, man aged 37) have Not detected annotations. Their metadata are preserved in donor IHC. These live-source records establish the sample mismatch; they do not replace the pinned v23 numerical data. Current portal RNA values differ slightly (13.1/0/0/6.9) and are not silently mixed into this run.
- Result: Variable, excluded from the retained set. This is an unresolved donor-pattern exclusion, not a conclusion that CTAG2 has caused cardiac injury.
Before reconsideration: identify the expressing cell types and isoforms, obtain paired RNA/protein evidence in RNA-positive and RNA-zero specimens, investigate donor/physiological/technical determinants, and independently validate any proposed subgroup. Additional relevant antigen-presentation evidence may inform that review. A new explicit review is required; clinical nomination never overrides the hold.
HPA's curated annotations outside testis are negative, but its primary antibody data include staining that HPA disregarded as presumed off-target. Keep the raw staining and curated interpretation separate. The donor rule here uses the pinned curated layer and does not pretend that every raw image is blank. Primary staining · Antibody validation · Heart evidence review.
Named targets
| Gene | Category | Donor review |
|---|---|---|
| CT83 | Retained | Not assessed |
| CTAG1A | Retained | Not assessed |
| CTAG1B | Retained | Not assessed |
| CTAG2 | Variable | Reviewed |
| CTCFL | Retained | Not assessed |
| GAGE1 | Retained | Not assessed |
| MAGEA12 | Excluded | Not assessed |
| MAGEA4 | Retained | Not assessed |
| PRAME | Retained | Not assessed |
| SAGE1 | Retained | Not assessed |
| SSX4 | Retained | Not assessed |
| SSX5 | Retained | Not assessed |
| XAGE1A | Retained | Not assessed |
| XAGE1B | Retained | Not assessed |
| ZNF730 | Excluded | Not assessed |
CTAG2 is the only loss from the previously retained 22-gene clinical/requested benchmark. The other 21 remain retained, including XAGE1A/B, CTAG1A/B, PRAME and MAGEA4. ZNF730 fails in cerebral cortex, bone marrow, lymph node and thymus. The 17 other removals come from the tighter selectivity rules. Every changed gene and reason · Clinical panel, including previously excluded targets.
The earlier frozen core/extended CTA APIs already exclude CTAG2. Historical reports and sequence exports retain their dated policies; the new explicit oncoref.cta_screen selection excludes Variable. This update does not certify reproductive-specific proteins or regenerate k-mer backgrounds.
Downloads and reproduction
- Gene results · Tissue evidence · Counts.
- Donor decisions · Pinned donor RNA · IHC donor metadata.
- Policy · Run manifest and checksums.
python scripts/cta_screen_report.py --output /tmp/cta-screen-audit
from oncoref import cta_screen
cta_screen.results() # full decisions, provenance and reasons
cta_screen.retained_gene_ids() # all 740 loci; excludes Variable
cta_screen.retained_gene_names() # known symbols; unnamed loci remain in IDs
cta_screen.variable_gene_names() # {"CTAG2"}
No donor model is fitted, no false-negative probabilities are invented, and an unreviewed gene is never labelled donor-negative. Human nTPM ceilings are not applied to dog/mouse measurements. API and limitations.