Predictive classification of metabolites from natural products holds immense promise for developing new therapeutics to treat diseases that involve cell-surface proteins. Cardiac glycosides and monoterpene indole alkaloids are metabolite compounds that can be extracted from plants and have therapeutic applicability in the treatment of irregular heartbeats and pain sensitivity through their interactions with ion pumps and neurotransmitter receptors. We seek to combine the results of analytical experiments involving ion mobility mass spectrometry (IM-MS) with computational studies involving open-source cheminformatics software. Our goal for this project is to predict the collision cross section (CCS) values of modeled structures and match with our IM-MS experimental results. Using a standard off-the-shelf cheminformatics software (RDKit), we can rapidly generate molecular ensembles based on simple SMILES strings and calculate different geometric properties that allow us to cluster these ensembles and obtain the most representative conformers to match with experiment. Using an open-source quantum chemistry software (Psi4), we can generate Mulliken charges of the representative conformers to conduct CCS calculations and compare with IM-MS experiments. We are currently applying this method to a suite of cardiac glycosides and monoterpene indole alkaloids to create a library of ion mobility data for rapid classification and identification of mixtures of plant-based toxin molecules. This combined experimental and computational approach will allow scientists to rapidly analyze and identify toxins in unknown mixtures based on their mobilities and CCS values.