Chemical space exploration with Molpher: Generating and assessing a glucocorticoid receptor ligand library

Computational exploration of chemical space is important in modern cheminformatics research as it can accelerate the discovery of new biologically active compounds. In this study, we present a detailed analysis of a chemical library generated by one of the available molecular generators, Molpher. A targeted library of potential glucocorticoid receptor (GR) ligands was generated, and its composition was compared with a reference library randomly sampling chemical space. A random forest was used to determine the biological activity of ligands, and its applicability domain, which is essential to consider when predicting the biological activity of newly designed compounds, was incorporated using conformal prediction. It was demonstrated that the GR ligand library is significantly enriched with GR ligands compared to the random library. Moreover, through prospective analysis, it was shown that Molpher could design compounds that were later experimentally verified as active on the GR. Finally, a set of 34 potential GR ligands was proposed.