Lib-INVENT with docking and RFXscore
A scaffold-decorating molecule generator is rewarded with Glide docking and a trained rescoring model, and its weights are updated toward high scorers.
The loop
- Proposes
Decorations (substituents) added to a factor Xa scaffold, written as SMILES by Lib-INVENT, a REINVENT variant
- Judged by
Glide XP docking, ligand efficiency, the RFXscore rescoring model, QED and a factor Xa QSAR model
- Steers
Reinforcement learning updates the generator's agent network; a diversity filter penalises repeats
Headline result
In a factor Xa case study, docking followed by RFXscore gave the best results, especially when diverse structures fitting a binding site were wanted; RFXscore plus the factor Xa model left the most compounds after MM/GBSA and other filters. RFXscore had a test r2 of 0.494 and MAE of 1.086.
What this does not prove
MM/GBSA was applied afterwards as an independent check, not as a training reward. Everything is computed binding, with no lab measurements, and the target is a single well-studied protein (factor Xa).
How it generalizes the harness
A SMILES generator is trained by a reward built from docking and a learned rescoring model rather than being prompted.
Classification
| What it proposes | Molecule |
|---|---|
| Generative model | Lib-INVENT RNN prior and agent (scaffold decoration) |
| Algorithmic authority | Glide XP docking, RFXscore rescoring, MM/GBSA post-check |
| Feedback | Reward signal |
| What gets updated | Model weights |
| When it runs | During training |
| Validity | Approximate physics (simulation, not lab results) |
| Evidence | Experiments reported |
| Lineage | Application |