Mapping EGFR resistance-bypass mechanisms across the oncology graph

Operator-authored prose pending

EGFR TKI resistance as an R&D problem: the third-generation osimertinib bypass tracks are mostly known, but the under-developed nodes (NRG1 fusions, splice-related effectors) move slowly between literature and pipelines — operator writes this section as long-form prose, ~150–250 words, footnoted, slow. Scaffold only.

Operator-authored prose pending

osimertinib's market position, the clinical sequencing reality, which bypass tracks each major pharma has internalised, where the gap is — operator writes this section as long-form prose, ~150–250 words, footnoted, slow. Scaffold only.

Operator-authored prose pending

interpret the figure: MET dominates (correct, expected); AXL is mid-tier and partly explored; NRG1 is the under-served third tier with strong literature signal but no clinical compound — operator writes this section as long-form prose, ~150–250 words, footnoted, slow. Scaffold only.

Operator-authored prose pending

resistance_relations edges from a deterministic keyword pass → bypass-target aggregation → drugability_gap composite (evidence × confidence ÷ patent count) → STRING augmentation for PPI-only bypass tier — operator writes this section as long-form prose, ~150–250 words, footnoted, slow. Scaffold only.

Operator-authored prose pending

portfolio implications: a small-team play is on NRG1 / HER3 fusion biology; the MET space is too crowded for new entrants but ripe for combination partnerships — operator writes this section as long-form prose, ~150–250 words, footnoted, slow. Scaffold only.

Operator-authored prose pending

resistance edges are keyword-mined (GLiREL has no resistance edge type yet); confidence is paper-level not assertion-level; indication filter is reserved (paper-indication links aren't modelled) — operator writes this section as long-form prose, ~150–250 words, footnoted, slow. Scaffold only.