In this two-part tutorial, we’ll show you how to use molSimplify to collect autocorrelation-based descriptors from molecular structures and use those to make predictions using a simple kernel ridge regression (KRR) model, as shown in our recent paper.