2025/08/20 by Pablo Arantes, Souvik Sinha, Giulia Palermo · 1 voice
Biochemistry, Genetics and Molecular Biology · #Chemical Synthesis and Analysis
paper · pdf · doi:10.26434/chemrxiv-2025-m26n9
In molecular studies, the accurate parametrization of small molecules stands as an essential yet growing demand. Addressing this, we introduce ParametrizANI, a tool crafted explicitly for establishing detailed protocols for dihedral parametrization us- ing GAFF and OpenFF force fields. The robust PyTorch-based program, TorchANI, forms the backbone of ParametrizANI, functioning as a benchmark to uphold pre- cision in parametrization tasks. TorchANI plays a pivotal role in facilitating the training and inference of ANI (ANAKIN-ME) deep learning models, which are fun- damental in predicting potential energy surfaces and a spectrum of other molecular system attributes. Our work on ParametrizANI goes beyond just creating a tool; it’s about building a research-friendly environment, free from the constraints of limited resources. We’re committed to democratizing research, enabling teams of all sizes to perform dihedral parametrization with DFT-level accuracy. This tool opens up new possibilities in molecular dynamics and related fields. It marks a significant step for- ward in improving the scientific community’s ability to parametrize small molecules. For a detailed look at its features, we invite you to check out ParametrizANI on GitHub (https://github.com/palermolab/ParametrizANI).