PACMAN

A Partial Atomic Charge Predicter for Porous Materials based on Graph Convolutional Neural Network (PACMAN)

Usage

from PACMANCharge import pmcharge
pmcharge.predict(cif_file="./test/Cu-BTC.cif",charge_type="DDEC6",digits=6,atom_type=True,neutral=True,keep_connect=True)
pmcharge.Energy(cif_file="./test/Cu-BTC.cif")

• cif_file: cif file (without partial atomic charges) [cif path]
• charge-type (default: DDE6): DDEC6, Bader, CM5 or REPEAT
• digits (default: 6): number of decimal places to print for partial atomic charges. ML models were trained on a 6-digit dataset.
• atom-type (default: True): keep the same partial atomic charge for the same atom types (based on the similarity of partial atomic charges up to 2 decimal places).
• neutral (default: True): keep the net charge is zero. We use "mean" method to neuralize the system where the excess charges are equally distributed across all atoms.
• keep_connect (default: True): retain the atomic and connection information (such as _atom_site_adp_type, bond) for the structure.

Website & Zenodo

PACMAN-APPlink
github repositorylink

Reference

@article{doi : 10.1021/acs.jctc.4c00434 ,
        author = {Zhao, Guobin and Chung, Yongchul G.},
        title = {PACMAN: A Robust Partial Atomic Charge Predicter for Nanoporous Materials Based on Crystal Graph Convolution Networks},
        journal = {Journal of Chemical Theory and Computation},
        volume = {20},
        number = {12},
        pages = {5368-5380},
        year = {2024},
        doi = {10.1021/acs.jctc.4c00434},
        note ={PMID: 38822793},
        URL = {https://doi.org/10.1021/acs.jctc.4c00434},
        eprint = {https://doi.org/10.1021/acs.jctc.4c00434}
        }

Bugs

If you encounter any problem during using PACMAN, please email sxmzhaogb@gmail.com.