A machine learning-based chemoproteomic approach to identify drug targets and binding sites in complex proteomes

Abstract

Chemoproteomics is a key technology to characterize the mode of action of drugs, as it directly identifies the protein targets of bioactive compounds and aids in the development of optimized small-molecule compounds. Current approaches cannot identify the protein targets of a compound and also detect the interaction surfaces between ligands and protein targets without prior labeling or modification. To address this limitation, we here develop LiP-Quant, a drug target deconvolution pipeline based on limited proteolysis coupled with mass spectrometry that works across species, including in human cells. We use machine learning to discern features indicative of drug binding and integrate them into a single score to identify protein targets of small molecules and approximate their binding sites. We demonstrate drug target identification across compound classes, including drugs targeting kinases, phosphatases and membrane proteins. LiP-Quant estimates the half maximal effective concentration of compound binding sites in whole cell lysates, correctly discriminating drug binding to homologous proteins and identifying the so far unknown targets of a fungicide research compound.

Challenges and advances in clinical applications of mesenchymal stromal cells

“…AI can find new molecular compounds and emerging drug targets much faster than traditional methods, thus speeding up the progress of drug development [ 184 , 185 ]. At the same time, AI can more accurately predict the follow-up experimental results of new drugs, so as to improve the accuracy at each stage of drug development [ 186 ].…”

Machine learning in clinical decision making

“…ML can be utilized to discern features indicative of drug-target binding. 43 One recent ML breakthrough, which may greatly facilitate drug discovery, is an AI pipeline developed by Google, predicting 3D protein structures from amino acid sequences. 44 This impressive development may enable to generate predictive knowledge on active sites within the tertiary structures of proteins, thereby facilitating the design of small-molecule inhibitors.…”

Proteomic Approaches to Study SARS-CoV-2 Biology and COVID-19 Pathology

“… 173 While the initial LiP-SMap identified 37 putative drug targets for cells treated with Rapamycin, the additional LiP-Quant scoring method could confirm FKBP1A as the highest-ranking candidate protein target. 173 As of this publication, there have been no published studies using LiP-MS to study SARS-CoV-2 or any recently identified drugs that are in or being considered for clinical trials. However, in light of the ongoing large scale drug discovery efforts for COVID-19, the relatively simple experimental design of LiP-MS and its broad applicability make it an ideal technique to identify cellular targets of existing drugs in clinical trials or for prioritizing drugs or antibodies based on their off-target reactivity.…”