Machine learning classification can reduce false positives in structure-based virtual screening

Abstract

With the recent explosion in the size of libraries available for screening, virtual screening is positioned to assume a more prominent role in early drug discovery's search for active chemical matter.Modern virtual screening methods are still, however, plagued with high false positive rates: typically, only about 12% of the top-scoring compounds actually show activity when tested in biochemical assays.We argue that most scoring functions used for this task have been developed with insufficient thoughtfulness into the datasets on which they are trained and tested, leading to overly simplistic models and/or overtraining. These problems are compounded in the literature because none of the studies reporting new scoring methods have validated their model prospectively within the same study. Here, we report a new strategy for building a training dataset (D-COID) that aims to generate highly-compelling decoy complexes that are individually matched to available active complexes. Using this dataset, we train a general-purpose classifier for virtual screening (vScreenML) that is built on the XGBoost framework of gradient-boosted decision trees. In retrospective benchmarks, our new classifier shows outstanding performance relative to other scoring functions. We additionally evaluate the classifier in a prospective context, by screening for new acetylcholinesterase inhibitors. Remarkably, we find that nearly all compounds selected by vScreenML show detectable activity at 50 µM, with 10 of 23 providing greater than 50% inhibition at this concentration. Without any medicinal chemistry optimization, the most potent hit from this initial screen has an IC50 of 280 nM, corresponding to a Ki value of 173 nM. These results support using the D-COID strategy for training classifiers in other computational biology tasks, and for vScreenML in virtual screening campaigns against other protein targets. Both D-COID and vScreenML are freely distributed to facilitate such efforts.

Machine learning classification can reduce false positives in structure-based virtual screening

“…The D-COID dataset is available at https://data.mendeley.com/datasets/8czn4rxz68/ ( 93 ). vScreenML is available at https://github.com/karanicolaslab/vScreenML .…”

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]. Computer-aided drug design techniques are thus revolutionizing MSCs therapies.…”

Machine‐learning scoring functions for structure‐based virtual screening

“…Adeshina et al 35 presented an unusually comprehensive study. Not only they investigated a new ML‐based SF (vScreenML), they also presented a new benchmark (D‐COID) and a prospective application of vScreenML, the two latter parts analyzed elsewhere in this review.…”

“…Other ways to make benchmarks more realistic have been proposed such as considering true instead of assumed inactives 110 or selecting decoys that are 3D similar to their actives. 35 Benchmarks should be compared to selecting these assumed inactives at random and ultimately assessing how well these anticipate prospective performance. The latter was done by Sun et al, 52 who…”

“…So how can we effectively anticipate which SFs would work best on a given target? Evaluating the SFs on a set of actives and their decoys selected in a different manner than those employed in the training set is a way to ensure that we are not exploiting the way these molecules were selected 35,37,63,77,78 . In addition, some authors have proposed ways to make benchmarks more realistic by considering true instead of assumed inactives 110 or selecting decoys that are 3D similar to their actives 35 .…”