The PubChemQC Project

NEWS

Please use The Public Computational Chemistry Database Project, our web interface to pubchemqc dataset.

We provide quantum chemical results for molecules of the PubChem Project: input, output, and mol files. The molecular geometires in the input files and the mol files are already optimized, you may also want to check by output files, or recalculation.

Density Functional Theory and B3LYP (Becke + Slater + HF exchange, with LYP + VWN1RPA e- gas formula) functional.
6-31G(d) (aka 6-31G*) basis set.
Geometry optimization
Excited state calculation by time depended density functional theory, using the geometry obtained above with 6-31+G(d) basis set.
GAMESS (Linux version)

Get molecular information from the PubChem Project
Extract Canonical SMILES description for each molecule by OpenBABEL.
Obtain initial geometories as 3Dify by "--gen3d -h" by OpenBABEL.
Perform PM3 geometry optimization calulation by GAMESS.
Extract PM3 optimized geometry and perform Hartree-Fock STO-6G geometry optimization calulation by GAMESS.
Extract Hatree-Fock STO-6G optmized geometry and perform DFT B3LYP geometry optimization calulation by Firefly.
Extract DFT B3LYP optmized geometry and perform DFT B3LYP geometry optimization again by GAMESS with tighter threshold.
Extract DFT B3LYP geometry and perform DFT B3LYP geometry optimization calulation again for sure, and reduce confusions for input files we provide.
Perform TD-DFT calulation with 6-31+G(d) for optimized geometry using B3LYP/6-31G(d), and obtain 10 lowest excited states (we added diffuse function, as excited states may contain diffuse orbitals).
Upload TD-DFT/DFT B3LYP input/output files.
Upload the mol files, too. This is for GaussView (and/or Gaussian) users and other quantum chemical program package users. In any case, you can use OpenBABEL to convert to the formats which your favorite program packages employ.
Upload results daily.
Almost everything has been automated ;-)

The results are in as is basis. We are not sure results are correct or not.
The PubChem Project provides over 75,000,000 molecules which are far beyond from our calculation capacity. (this is our challenge!)
The molecular weight should be less than 500.
Only for the singlet and neutral molecules.
Only for molecules which contain H, He, Li, Be, B, C, N, O, F, Ne, Na, Mg, Al, Si, P, S, Cl, Ar, K, Ca, Sc, Ti, V, Cr, Mn, Te, Co, Ni, Cu, Zn (basis set limitation of PM3 method, 6-31G(d) and STO-6G basis).

Politically, you must agree to uploading results to this project as well as the PubChem Project.
You need machine(s) to run GAMESS .
Then ask me ;-)

2016/8/23 : Add molecule geometry initial guess generator.
2016/8/22 : Add molecule query.
2016/1/7: Our paper has been published The PubChemQC project: A large chemical database from the first principle calculations, Nakata Maho, AIP Conf. Proc. 1702, 090058 (2015).
2015/5/27 : Over 2 million molecules (2122146 molecules (excited states are available for 2016173 molecules) last updated : Wed May 27 02:39:30 JST 2015)
2015/5/4 : Our paper has been accepted by CC symposium in ICCMSE 2015.
2014/4/24 : Add excited states by TDDFT calculation.
2014/4/4 : This project has been moved to RIKEN, compress log files by xz with 50,000 molecules!
2013/12/21 : This project has been launched.