The PubChemQC Project: a large chemical database from the first principle calculations

Maho Nakata
DOI: https://doi.org/10.48550/arXiv.1512.08572
IF: 2.552
2015-12-29
Chemical Physics
Abstract:In this research, we have been constructing a large database of molecules by {\it ab initio} calculations. Currently, we have over 1.53 million entries of 6-31G* B3LYP optimized geometries and ten excited states by 6-31+G* TDDFT calculations. To calculate molecules, we only refer the InChI (International Chemical Identifier) representation of chemical formula by the International Union of Pure and Applied Chemistry (IUPAC), thus, no reference to experimental data. These results are open to public at http://pubchemqc.riken.jp/. The molecular data have been taken from the PubChem Project (http://pubchem.ncbi.nlm.nih.gov/) which is one of the largest in the world (approximately 63 million molecules are listed) and free (public domain) database. Our final goal is, using these data, to develop a molecular search engine or molecular expert system to find molecules which have desired properties.
What problem does this paper attempt to address?