Check out our latest publication Predicting Polypharmacology with Web-Based Tools in Wiley!

Abstract

Identifying possible off‐targets is critical during drug development. Estimating polypharmacology can be nowadays performed online thanks to the availability of various web‐based target prediction tools. In this chapter, we review eight such tools that are currently active, detailing their source data and the molecular encoding and statistical models used to assess the potential polypharmacology of a query compound. The best‐performing target assignment method across these different tools consists in comparing the query compound with molecules of known activity, taken from a reference dataset such as ChEMBL, relying on a molecular fingerprint, generally the Morgan fingerprint, to encode molecular structure, and Naïve Bayes statistics using individual bits of the molecular fingerprint as independent information sources to perform the prediction.

Author(s) Maedeh Darsaraee, Sacha Javor, and Jean-Louis Reymond