Insilico Medicine Reports AI-Designed PROTAC That Degrades and Inhibits PKMYT1
This month, Insilico Medicine has published a paper in Nature Communications describing an AI-generated PROTAC that targets PKMYT1, a kinase linked to synthetic lethality in cancers with CCNE1 amplification or mutations in FBXW7 and PPP2R1A. Insilico presents the work as evidence that its generative chemistry platform can move from standard small-molecule inhibitor design into more complex, modality-hybrid protein-degrading drugs such as PROTACs.
The new molecule, called D16-M1P2, was created with Insilico’s Chemistry42 platform, which designed both the inhibitor ‘warhead’ and the linker that connects it to the E3 ligase binder, and Insilico describes D16-M1P2 as a first-in-class PKMYT1 degrader generated by this platform.
The team reports that the PROTAC works in two ways: it degrades PKMYT1 at lower doses and inhibits its residual activity at higher doses. In preclinical studies, the compound showed selective PKMYT1 targeting, sustained pathway suppression, and anti-tumor effects in biomarker-defined models, including subsets of breast and bile duct cancers. The molecule also showed favorable oral bioavailability in animal studies.
According to the authors, this dual mechanism is needed because degradation alone left some PKMYT1 signaling active in preclinical tests, so adding direct inhibition is meant to shut the pathway down more completely and address selectivity, resistance, and non-catalytic functions.
The PROTAC has advanced to pre-candidate validation, building on an earlier Journal of Medicinal Chemistry paper from February 2025 that detailed the design path for the optimized PKMYT1 inhibitor scaffold used here.
Insilico’s broader oncology portfolio has also produced two recent studies. One nominates ISM3830, a CBLB-targeting compound for tumor immunotherapy, and another reports the design and evaluation of DGKα inhibitors for cancer-associated immune modulation.
The publication arrives shortly after Insilico announced eight AI-designed oral cardiometabolic programs at BIO-Europe, showing the range of outputs produced by the same AI platform—from orally available small molecules to complex, multi-component degraders. New candidates range from GLP-1 receptor agonists to NR3C1, NLRP3, GIPR, APJ, DACRA, and Lp(a) programs. These molecules span lead identification through IND-enabling work and were described as designed through multi-parameter generative optimization aimed at oral dosing, longer half-life, and combination-use profiles.
Pharma.AI also underpinned a new research and licensing agreement with Eli Lilly, in which Insilico will apply its generative design and optimization tools to jointly defined targets. The deal includes more than $100 million in potential payments and expands an earlier software relationship established in 2023, when Lilly first began using components of Insilico’s AI platform in discovery workflows.
Topic: AI in Bio
