AI Model RetroChimera Speeds Up Molecule Synthesis with High Accuracy

Researchers from Microsoft Research and collaborators have developed an AI model called RetroChimera to help automate the process of molecule synthesis. The system, described in a study published in Nature, uses multiple models with complementary strengths to propose promising synthesis pathways for target molecules.

The development of RetroChimera builds on Microsoft's longstanding research into using AI to accelerate scientific discovery, including advances in chemistry and drug development. The researchers analyzed common shortcomings in existing retrosynthesis systems, including difficulties incorporating less frequent but strategically important reactions and a tendency to generate inaccurate predictions.

According to the study, RetroChimera was able to propose fully accepted sequences of reactions for nine molecules selected to benchmark its performance against other models. When expert chemists evaluated proposed pathways for these molecules, they preferred RetroChimera's approach about 64% of the time.