READRetro

READRetro (Retrieval-Augmented Dual-view Retrosynthesis) is a practical bio-retrosynthesis tool designed to predict the biosynthetic pathways of plant natural products. Predicting these pathways is challenging but crucial for understanding plant ecology and biochemistry. READRetro integrates deep learning models and database retrieval methods to propose plausible biosynthetic routes from target molecules to metabolic building blocks.

Check out the User Guide section for further information, including how to installation the project.

Overview of READRetro Model Architecture