Summary

The idea

The electronic properties of molecules can be greatly influenced by their accessible conformation. As a result, it is essential to have a proper tool to explore the conformation space of small molecules. Although conformation is usually explored at a classical level using force fields like UFF or MMFF94, these lack descriptions of quantum effects. Conversely, starting from scratch at a high QM level of theory is not advisable due to two main reasons: the initial structure is crucial to ensure convergence, and these methods are computationally expensive.

An intermediate level of theory, which is accessible for almost any current computer, is the semi-empirical level of theory. aleimi attempts to automate the combination of classical, semi-empirical, and QM level theories for the conformational analysis of small molecules.

aleimi uses RDKit for stochastic generation of conformer and classical optimization, MOPAC for the semi-empirical part. And Psi4, Gaussian or Orca for the QM simulations. For the last part, aleimi only generate the files to launch on the cluster.

General workflow

Here we show an example workflow of aleimi that can be obtained with the following parameter file:

# aleimi.confgen.main
numConfs: 1000
rdkit_d_RMSD: 0.2
UFF: false
mopac_keywords: PM7 precise ef xyz geo-ok t=3h THREADS = 2 # You can also add `EPS=78.4` to take into account, implicitly, the solvent
# aleimi.boltzmann.main
Bd_rmsd: 1.0
Bd_E: 0.0
# aleimi.extractor.main
energy_cut: 2
engine: psi4