Precision fragment addition: domain-specific DeepFrag2 models for smarter lead optimization

Key Points

• The aim is to improve lead optimization techniques using DeepFrag2, a machine-learning tool focused on fragment addition.
• Utilization of the DeepFrag2 model for lead optimization through fragment addition.
• Training the model on fragments of varying sizes and charge states.
• Fine-tuning the model on specific receptor classes to enhance accuracy.
• DeepFrag2 demonstrates higher accuracy when trained on domain-specific fragment parameters.
• Performance improves significantly with the adaptation to specific receptor classes.