Cybergenic
All findings

KRAS + INPP4B

KRAS and INPP4B co-occurrence in non-small cell lung cancer (MSK-CHORD)

Strengthened: held up under a stratified controlCo-occurrence · Evidence: pattern across patients, not yet tested within subtypes

DataMSK-CHORD (2024) · Non-small cell lung cancer · 7,718 tumours · both altered in 22 (11 expected if independent)

Engine's conclusion

Strengthened. Significant, with a real effect in a consistent direction, and it still holds when tumours are compared only within the same tumour type (and within the same molecular subtype, wherever this cancer's subtypes are known in the data). No artifact check failed.

On the map: An early signal. Neither its statistical strength nor its expected value is high enough yet to count as a discovery.

Not called novel: a PubTator3 search finds 1 paper naming both KRAS and INPP4B in non-small cell lung cancer in the title or abstract. See the PubTator3 search.

Statistical support49%Significance (FDR q)0.0016Signal (x)14%Efficacy (y)2%

Signal (x) is the corrected significance on the map's scale, discounted from 30% to 14% because the cohort-composition confound is not yet controlled. Efficacy (y) is credibility times actionability.

Structure & site3D structure: AlphaFold

These genes are altered together more often than chance would predict.

Primary site

Lung

This finding's tumour type arises in the highlighted organ.

Drag to rotate. Dark marks the hotspot residues on the predicted fold. For a fuller read of this same model, open the interactive structure workbench for this protein on Nanoverse (a companion site we also run), which maps its binding pockets, per-region confidence, and the likely effect of a mutation.

KRASOncogeneApproved therapy

KRAS proto-oncogene GTPase

Oncogene: when altered it gets stuck ON and drives growth.

Central RAS signal transducer. G12C is now druggable (sotorasib, adagrasib); G12D inhibitors are emerging.

Pathway
RTK to RAS to MAPK
Hotspots in 3D
G12 · G13 · Q61 · A146
Drives
Pancreatic, Colorectal, NSCLC
KRAS mutations in non-small cell lung cancer →KRAS co-mutations by tumour type →Other findings with KRAS →

No structure profile is curated for this gene.

INPP4BGene

A human gene in this finding. Not in the curated driver set, so no structured profile is shown.

Other findings with INPP4B →

Anatomical figure from the EBI Expression Atlas anatomogram, and predicted structures from the AlphaFold Protein Structure Database (both CC-BY-4.0). A model shows the protein's healthy fold; a highlighted residue marks a recurrent cancer position, not a prediction of a specific mutation's effect.

Statistical evidence
Test p (raw)
1.5 × 10-6
Effect (log₂ odds ratio)
1.8
Papers
183
Abstracts naming both
1
Trials
0
Contingency table: the source of the q7,718 tumours

The raw 2×2 the FDR q and log₂ odds ratio above were computed from. It is the immutable evidence the score re-derives from: every number on this page traces back to these four counts.

Gene stateKRAS alteredKRAS wild-type
INPP4B altered2217
INPP4B wild-type2,1025,577

both 22 (expected 11 if independent) · INPP4B-only 17 · KRAS-only 2,102 · neither 5,577

Sources
cBioPortal / MSK-CHORD (2024)Genome NexusOmniPath / Reactome / STRINGUCSC XenaOpen TargetsPubTator3NIH iCiteEurope PMCClinicalTrials.gov
Anytime-valid evidence (e-value)e = 12

Strong evidence. Below the α=0.05 reject threshold (e≥20). Unlike a single-look p, this stays valid under the engine's continuous re-testing.

Why it sits here on the map

Molecular-subtype certification is unresolved. Recorded lineage only. Within-lineage molecular subtype remains unresolved. The association remains visible, but does not qualify as a subtype-controlled discovery.

Opacity: novelty (display only)31%

The two axes are variables: signal (x) and efficacy (y = credibility × actionability). A point high on both lands in the frontier (top-right). A strong signal with low credibility sits low and to the right: statistically strong, not worth pursuing.

KRASOncogeneINPP4BSotorasibAdagrasib
HypothesisModel-generated hypothesis (not evidence)
Written by gpt-6-luna

A language model wrote this from the evidence listed below. It is annotation: the engine's verdict above comes only from the measured statistics, never from this text.

The most likely explanation is unresolved molecular-subtype or lineage composition: KRAS- and INPP4B-altered tumours may be overrepresented in the same NSCLC subtype(s), producing population-level co-occurrence that will attenuate after subtype adjustment. Provisional ranking: (1) subtype/lineage composition, (2) cohort-specific data artifact or winner's curse, (3) a shared upstream driver, (4) co-amplification of neighbouring genes on one amplicon, and (5) genuine within-tumour cooperation.

Argument strength (model self-report)18%
After adversarial review18%
The argument
  1. 1The association is statistically notable and has a substantial estimated effect: burden-aware testing gives q = 1.6e-3, and the within-lineage log2 odds ratio is 1.8, with 22 tumours carrying both alterations. This supports investigating the pattern, but does not identify its cause.E1E18
  2. 2Subtype composition is the leading explanation because the adjustment records lineage but leaves molecular subtype unresolved; the measured heterogeneity test does not establish that the effect is homogeneous. A subtype-stratified association could therefore remain behind the reported lineage-adjusted signal.E2
  3. 3The strongest deflating alternative is cohort-specific signal or a winner's-curse/data artifact: the association has not recurred in another independent patient source, despite surviving correction in this cohort.E3E18
  4. 4A shared upstream driver is possible but not specifically supported by these data. Co-amplification is also unsupported: the evidence provides no copy-number or shared-amplicon result. Neither explanation is as directly motivated as the unresolved subtype confounding.E1E2E4
  5. 5The supplied paper abstracts do not establish a KRAS–INPP4B relationship: the colorectal-cancer abstract concerns AhR and high-fat-diet progression, while the NSCLC abstract concerns acquired BRAF-AGK fusion after therapy. Neither studies this gene pair, so the within-tumour cooperation reading is unearned.E10E11
  6. 6This lead PARTIALLY CONVERGES with the corpus's NSCLC warning that unresolved subtype composition may explain a KRAS-associated population pattern; the specific KRAS–INPP4B association remains a distinct lead. The colorectal and pancreatic KRAS findings concern different settings and gene pairs, so they do not directly validate this explanation.E14E15E16E17
What would confirm or refute it

Test the subtype-composition hypothesis in an independent, adequately sized NSCLC cohort with harmonized molecular-subtype and histology annotations: estimate the KRAS–INPP4B association before and after adjustment or stratification for subtype, lineage, and relevant technical factors, and report the adjusted odds ratio with uncertainty. Attenuation toward the null across subtype-adjusted analyses would support this explanation. To earn a within-tumour cooperation claim instead, perform a clonal or single-cell assay showing both alterations in the same tumour cells, followed by a functional co-dependency test.

Strongest competing explanation (model)

This is a screened association from only one independent patient source, and subtype remains unresolved, so confidence is limited despite the FDR-significant signal. Cohort composition or a winner's-curse artifact could account for it; shared upstream biology and co-amplification are not demonstrated. No supplied paper actually studies the KRAS–INPP4B pair, which strengthens the claim that a specific within-tumour mechanism remains under-explored rather than supporting one.

Adversarial review (second model pass)

Weakest link: Subtype or lineage composition is an untested explanation rather than a result of the evidence: subtype remains unresolved, and the association has not replicated beyond one cohort, leaving cohort-specific artifact or winner's curse plausible.

Cross-cohort agreement (intra-snapshot)single source

Significant (after FDR correction) in one independent source of this tumour type so far. The engine records where a pattern is significant, not every source that could test it, so this calls for caution: a pattern selected from many tests tends to overstate its effect. It does not count against the pattern.

Evidence the argument cites

Model-written annotation, not a finding. The statistics it cites are real and are shown above; the argument connecting them is a hypothesis that needs wet-lab or clinical validation before it means anything.

Cite this finding
Permalink
https://cybergenic.im/findings/8aa954c6-11d6-46d1-b7f1-fc698e242944
First seen
29 Sep 2026
Prediction registry
Locked 29 Sep 2026, row hash af9752fd3bbc94c68827ed24045b2630 (not testable out of sample). How the registry works
BibTeX
@misc{cybergenic_8aa954c6,
  author       = {{Cybergenic}},
  title        = {{KRAS and INPP4B co-occurrence in non-small cell lung cancer (MSK-CHORD)}},
  year         = {2026},
  howpublished = {\url{https://cybergenic.im/findings/8aa954c6-11d6-46d1-b7f1-fc698e242944}},
  note         = {Research engine finding, first seen 2026-09-29; prediction locked 2026-09-29, registry row hash af9752fd3bbc94c68827ed24045b2630},
  month        = sep
}
See it on the live map Mutations in non-small cell lung cancerINPP4B and KRAS in every cohort

Frontier alerts