A new NPJ Precision Oncology paper from NYU Langone Health, in collaboration with Deep Origin’s computational biology team, maps how lung cancers might outsmart the latest wave of targeted EGFR inhibitors and shows our tools can provide biological insights.

Using deep mutational scanning—a high-throughput technique that tests tens of thousands of possible amino-acid substitutions in parallel—the team at Langone built an exhaustive resistance atlas for BLU-945, a fourth-generation tyrosine kinase inhibitor designed to overcome the stubborn T790M and C797S mutations that limit earlier EGFR inhibitors.

The screen recovered known osimertinib escape routes (like L718Q/V) and revealed new liabilities at residues including Q791L/K, K716T, and K728E. Computational simulations by Deep Origin scientists showed how these subtle changes reshape the drug’s binding pocket and weaken inhibitor affinity. These predicted resistance mutations were later found in clinical cases, where patients acquiring L718X mutations experienced early progression on BLU-945.

Why does this matter for drug discovery scientists?

Deep mutational scanning provides a comprehensive list of mutations that can confer resistance to drugs, but often provides limited information bridging to structural mechanisms of resistance. Using our molecular modeling tools, Deep Origin's team proposed a change in molecular structure that explains resistance, offering drughunters a path to developing new drugs that overcome this resistance mechanism.