Design of a novel, potent CDC7 inhibitor development candidate with high ligand efficiency and optimized properties

CDC7 is a cell-cycle kinase target for hematological cancers and solid tumors, but developing highly potent (ideally picomolar), substrate-competitive inhibitors with balanced selectivity and drug-like properties had eluded the field for over 20 years. The team used LiveDesign for crowdsourced collaboration among 14 researchers, FEP+ for binding-affinity prediction accurate to within 1.0 kcal/mol, IFD-MD for off-target binding predictions, and membrane permeability modeling to address cell-potency gaps. They scored 24,000 compounds computationally and synthesized 226 total compounds, reaching a development candidate in 25 months: SGR-2921, with ~10 pM binding potency, ligand efficiency above 0.6, 10,000x selectivity against an identified off-target, and strong anti-tumor activity in AML xenograft models and patient-derived samples.

The customer is anonymised in Schrödinger’s own case study — not withheld by us.

Schrödinger is a physics-plus-machine-learning computational platform for molecular design; all top-20 pharma companies by revenue are customers of its drug-discovery software.

This sits in the Build stage of the AI use-case lifecycle: Named examples from the foundation-model labs and platforms pharma teams use to actually construct, train and generate the model or application.