Deep Ranking Analysis by Power Eigenvectors (DRAPE): A polypharmacology case study

Multi-criteria decision making processes comprehend several ranking methods able to handle multiple and often conflicting criteria and sorting objects according to a definition of the optimality direction for each criterion. In the field of polypharmacology it can be useful for virtual screening or prioritization purposes to rank molecules distinguishing the most multi-target ones, i.e. molecules able to interact with different targets, from the selective or inactive ones. In this work, the Deep Ranking Analysis by Power Eigenvectors approach is applied to a small set of molecules characterized by their half maximal binding concentrations for seven different nuclear receptors. The existence of a correspondence between the DRAPE ranking and a manually grouping of molecules based on their multi-target behaviour was verified. Moreover, a comparison between DRAPE rankings and those obtained from five traditional methods was carried out in order to highlight the main similarities and the differences of the approaches.