Accelerating drug target inhibitor discovery with a deep generative foundation model
Problem
Discovering small-molecule inhibitors for emerging drug targets when target structure or known active molecules are unavailable.
Why it matters
Faster inhibitor discovery could speed treatments for emerging diseases such as SARS-CoV-2.
What the AI did
A deep generative foundation model trained on protein sequences, small molecules and their interactions designed inhibitor candidates from target sequence alone; candidates were synthesized and experimentally tested.
Results
Micromolar-level in vitro inhibition for 2 of 4 synthesized candidates per target; most potent spike RBD inhibitor active against several variants in live virus neutralization assays.
Impact
Experimentally validated inhibitors for two SARS-CoV-2 targets, but not a clinical treatment.
Record
| Venue | Science Advances |
|---|---|
| Published | 2023 |
| Work type | article |
| Authors | Vijil Chenthamarakshan, Samuel C. Hoffman, David Owen, Petra Lukacik, Claire M. Strain-Damerell, D. Fearon, Tika R. Malla, Anthony Tumber et al. |
| Citations | 51 (OpenAlex, at import time) |
| References | 90 |
| Model or method | deep generative foundation model (protein sequence-conditioned) |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1126/sciadv.adg7865 |
| Open access | gold |
| Keywords | drug discovery, generative model, SARS-CoV-2, inhibitors, deep learning, experimental validation |
Importance rating
major — millions of lives, an industry or a field
Addresses inhibitor discovery for emerging drug targets, validated on SARS-CoV-2 spike RBD and main protease with live-virus neutralization; broad pandemic-relevant impact.