Microsoft Research has published and open-sourced RetroChimera, a new framework for retrosynthesis prediction that combines two complementary models to propose high-quality synthesis routes. The system integrates R-SMILES 2, a Transformer-based de-novo model, with NeuralLoc, a graph neural network (GNN) based model, using a learned ensembling strategy to rank predictions.

  • RetroChimera leverages the flexibility of R-SMILES 2 and the grounded accuracy of NeuralLoc to capture complementary reaction patterns.
  • In blind tests, PhD-level chemists preferred RetroChimera’s individual reaction predictions over preceding models and recorded literature reactions.
  • The model achieved a 90% acceptance rate for complete synthesis routes by experts, compared to 20-50% for its sub-models or baselines.
  • Implementation and weights are available on GitHub under the MIT license and via Microsoft Foundry.

RetroChimera aims to help researchers identify promising synthesis routes more efficiently, supporting faster design-make-test cycles in drug discovery and smart materials.