Descriptor Generation

In this example we are going to generate a collection of xtb-derived chemical descriptors as well as a collection of RDKit-derived descriptors. We are going to store them in .json format for each molecule. And we are going to create a csv file with the boltzmann averaged values of the descriptors that we have calculated per each molecule. The following scheme summarizes the contents of this example.


We are starting from a 'test.csv' file containing the SMILES of the molecules whose chemical descriptors we are going to calculate:


In this case we are going to start by generating some conformers of these molecules using rdkit (for more details on the conformer generation please check the Conformer Search section).

python -m aqme --csearch --input test.csv --program rdkit

Next we proceed to generate the descriptors which is fully automated by the QDESCP module.

python -m aqme --qdescp --program xtb --files "CSEARCH/*.sdf"