PTEN

The bull case

1. 1.

   A major, hotspot-driven cancer lesion. AlphaMissense flags 72.25% of 7,657 scored variants as pathogenic, with a hotspot at residue 127 (mean score 1.0) — PTEN loss/mutation is a well-characterized, high-impact event. (A biomarker strength, not a drug-target strength.)

   confidence: high·genetic_validation.variant_pathogenicity

2. 2.

   Sustained interest despite no asset. 100 patents across 214 organizations, with filings holding (24 in 2022, 24 in 2024). The field is actively working the target even without clinical chemical matter.

   confidence: moderate·competitive_landscape

3. 3.

   Deep literature. 257 curated papers support a well-evidenced view.

   confidence: high·coverage.dimensions.literature

The bear case

1. 1.

   Wrong target shape for inhibition — the defining fact. DepMap mean gene effect is +0.3956 (positive); PTEN is essential in only 1.4% of 1,208 models. Knocking PTEN out improves fitness — it is a loss-of-function tumor suppressor. There is no catalytic dependency to inhibit, and “restoring” a tumor suppressor is not a modality this graph evidences.

   confidence: high·genetic_validation, sign-read

2. 2.

   No clinical chemical matter. Max clinical phase 0; only 2 compounds (both preclinical); ChEMBL lists just 5 bioactivities. The target is essentially undrugged by direct means.

   confidence: high·competitive_landscape / coverage

3. 3.

   Compound coverage too thin to assess. 2 compounds — SAR, selectivity, and series competition are not visible in this graph.

   confidence: flagged·coverage.dimensions.compounds

4. 4.

   The PI3K-axis link is consensus and crowded. The one independently-evidenced downstream node, PIK3CA (PTEN loss → PI3K activation; 2 papers), carries 353 patents and phase-4 chemical matter — heavily worked, not open.

   confidence: moderate·resistance_bypass_map

The non-consensus angleNo strong angle

No strong non-consensus signal in current data; PTEN's profile largely matches its reputation as a loss-of-function tumor suppressor. No DepMap co-essentiality fired and there is no independently multi-sourced whitespace node — manufacturing an angle here would be reaching.

• A coherent PTEN interactome, but PPI-proxy only. The functional-coupling whitespace tool returns MAGI2, DLG1, PTK2 (FAK), MAST2, PREX2, SPOP, PIK3R2, NEDD4 (PPI 0.98+) — real PTEN regulators/partners. None is in Mosaic's curated target set, so the graph has no patent/compound coverage for them; their competitive status is not assessed (not a whitespace claim). This is PPI-proxy coupling, not DepMap co-essentiality, so it is not advanced. (Archetype caveat: an earlier unattended run phrased these as “a PTEN inhibitor + partner” — nonsensical for a tumor suppressor; the tumor-suppressor rule now phrases them as synthetic-lethal partners of PTEN-loss.)

  confidence: PPI-proxy — weak·kg_native.synthetic_lethal_whitespace (PPI-proxy)

  Patent / compound status: not assessed (out of coverage). gene outside Mosaic curated coverage — competitive landscape is a validated lead to verify externally, not a whitespace claim.

• PIK3CA is the only evidenced axis node, and it is consensus. Well-known biology (PTEN loss activates PI3K), heavily patented — not whitespace.

  confidence: moderate·kg_native.resistance_bypass_map (curated)

What would change this thesis

• If DepMap co-essentiality evidences a synthetic-lethal partner of PTEN-loss (none fired here), a real non-consensus angle emerges — the canonical way to drug a tumor suppressor.
• If a PTEN-restoration or PTEN-stabilizing modality (e.g. blocking the NEDD4-class degradation the interactome hints at) enters the graph, the “undruggable” bear weakens.
• If DepMap read in a PTEN-loss-selected context shows a collateral dependency, Bear #1's “no dependency” reading gains a therapeutic handle.