Quantitative Polypharmacology Profiling Based on a Multifingerprint Similarity Predictive Approach

Abstract

We present a new quantitative ligand-based bioactivity prediction approach employing a multifingerprint similarity search algorithm, enabling the polypharmacological profiling of small molecules. Quantitative bioactivity predictions are made on the basis of the statistical distributions of multiple Tanimoto similarity θ values, calculated through 13 different molecular fingerprints, and of the variation of the measured biological activity, reported as ΔpIC50, for all of the ligands sharing a given protein drug target. The application data set comprises as much as 4241 protein drug targets as well as 418 485 ligands selected from ChEMBL (release 25) by employing a set of well-defined filtering rules. Several large internal and external validation studies were carried out to demonstrate the robustness and the predictive potential of the herein proposed method. Additional comparative studies, carried out on two freely available and well-known ligand–target prediction platforms, demonstrated the reliability of our proposed approach for accurate ligand–target matching. Moreover, two applicative cases were also discussed to practically describe how to use our predictive algorithm, which is freely available as a user-friendly web platform. The user can screen single or multiple queries at a time and retrieve the output as a terse html table or as a json file including all of the information concerning the explored similarities to obtain a deeper understanding of the results. High-throughput virtual reverse screening campaigns, allowing for a given query compound the quick detection of the potential drug target from a large collection of them, can be carried out in batch on demand.

TIRESIA: An eXplainable Artificial Intelligence Platform for Predicting Developmental Toxicity

“…Nevertheless, developmental toxicity is one the most important toxicological endpoints under European Union’s REACH (Registration, Evaluation, Authorization and Restriction of Chemicals) regulations and United States Environmental Protection Agency (EPA) requirements. Thus, the use and development of alternative methods as in vitro assays and in silico methods are strongly encouraged. − It is acknowledged that modeling the complex outcomes related to this endpoint is extremely difficult because the current experimental methods (i) are not universal, (ii) cannot provide suitable evaluations for certain properties, and (iii) are very much expensive and time demanding …”

MolPredictX: Online Biological Activity Predictions by Machine Learning Models

“…Another web server by the same group is e-MuSSEL, developed by Ciriaco et al, 2021. [30] This algorithm automated the quantitative prediction of pIC 50 values for 4241 protein targets. This tool uses data selected from ChEMBL with a different filtering rule and doesn't require any calibration set for tuning parameters once new ChEMBL updates become available.…”

“…These two algorithms were combined by the group to form PLATO, which is an easy-to-use polypharmacology predictive platform, designed for finding new putative protein drug targets and quantifying bioactivity affinity value. [29,30] 3D-QSAR developed by Wang et al, 2021, [31] performs the complete QSAR modeling procedure and provides activity functions by integrating the functions of molecular structure generation, alignment, and molecular interaction field (MIF) computing and results analysis to provide a onestop solution.…”

“…MuSSEL developed by Alberga et al., 2019, [29] is a multifingerprint similarity search algorithm which predicts putative drug targets for a given query molecule and returns a quantitative assessment of its bioactivity in terms of Ki or IC 50 values. Another web server by the same group is e‐MuSSEL, developed by Ciriaco et al., 2021 [30] . This algorithm automated the quantitative prediction of pIC 50 values for 4241 protein targets.…”

“…Since e‐MuSSEL is actively enlarges its basin of experimental data and is based on a new statistical‐based algorithm it is a quick and easy method for screening a large amount of data. These two algorithms were combined by the group to form PLATO, which is an easy‐to‐use polypharmacology predictive platform, designed for finding new putative protein drug targets and quantifying bioactivity affinity value [29,30] …”