Optimizing Molecular Glues Using Free Energy Perturbation and Cofolding Methods
Bibliographic record
Abstract
Molecular glues, a class of small molecules that induce protein-protein interactions, hold significant promise as a therapeutic modality, offering access to new biology unlocking new protein targets with enhanced specificity. Despite these advantages, their rational design and optimization remain challenging due to the dynamic nature of their interfacial binding sites and complex structure-activity relationships. Computational methods for binding affinity prediction are crucial for accelerating drug discovery, with Free Energy Perturbation (FEP) being a gold standard for accuracy at high computational cost, and co-folding models like Boltz-2 offering speed but with unproven accuracy on novel targets. This study presents the first comprehensive evaluation of FEP and Boltz-2 for predicting the binding affinity of molecular glues to protein complexes. We assessed 93 compounds across six diverse target/effector complexes, yielding 140 unique protein-compound measurements with experimentally validated ternary complex binding and degradation data. Our results demonstrate that FEP consistently outperforms Boltz-2 across all datasets in terms of correlation and RMSE, showing good absolute predictability with RMSE values within 0.3-1.25 kcal/mol as well as strong correlations. In contrast, Boltz-2 exhibited poor absolute predictability (RMSE over 3 kcal/mol in some cases) and generally poor or even negative correlations. While FEP is computationally more expensive, its accuracy makes it a valuable tool for molecular glue optimization. The poor performance of Boltz-2 suggests it is not suitable for high-throughput screening of molecular glues, highlighting the need for more accurate, high-throughput machine learning methods for pre-FEP screening. This work underscores the current capabilities and limitations of computational methods in the challenging field of molecular glue discovery.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.002 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".