Chemprop v2: An Efficient, Modular Machine Learning Package for Chemical Property Prediction
David Graff, Nathan K. Morgan, Jackson Burns, Anna C. Doner, Brian Li, Shih‐Cheng Li, Joel Manu, Angiras Menon, Hao‐Wei Pang, Haoyang Wu, Akshat Shirish Zalte, Jonathan W. Zheng, Connor W. Coley, William H. Green, Kevin P. Greenman
Abstract
Accurate prediction of molecular properties is essential for computational design in many areas of chemistry. Deep learning has been used in these prediction tasks for a wide variety of molecular properties, and the availability of user-friendly open-source software implementing such architectures has democratized access to these methods. chemprop is one of the most popular examples of such software in this field. It implements a directed message-passing neural network (D-MPNN) architecture, enabling end-to-end learning of molecular properties directly from molecular graphs without the need for handcrafted descriptors or fingerprints. The original chemprop release was intended for use primarily via a command line interface, rather than programmatic use via a Python API. As the field has evolved, the need for increased modularity and usability in Python-based workflows has become clear. We completed a ground-up rewrite of chemprop that addresses this need, providing improvements in speed, extensibility, and overall user experience. We have conducted extensive benchmarking to demonstrate algorithmic parity with the original implementation, while seeing improvements of about a factor of 2 in execution time and a factor of 3 in memory usage. chemprop v2 effectively scales to multiple GPUs, which enables the training of more and larger models. chemprop v2 also includes some new features. Extensive Jupyter notebook tutorials and new documentation for all major functionality were also added. chemprop v2 preserves the predictive accuracy of its predecessor and enhances modularity, speed, and usability, empowering researchers to pursue computational molecular design more effectively.