Regression-Based Active Learning for Accessible Acceleration of Ultra-Large Library Docking

Structure-based drug discovery is a process for both hit finding and optimization that relies on a validated three-dimensional model of a target biomolecule, used to rationalize the structure-function relationship for this particular target. An ultralarge virtual screening approach has emerged recently for rapid discovery of high-affinity hit compounds, but it requires substantial computational resources. This study shows that active learning with simple linear regression models can accelerate virtual screening, retrieving up to 90% of the top-1% of the docking hit list after docking just 10% of the ligands. The results demonstrate that it is unnecessary to use complex models, such as deep learning approaches, to predict the imprecise results of ligand docking with a low sampling depth. Furthermore, we explore active learning meta-parameters and find that constant batch size models with a simple ensembling method provide the best ligand retrieval rate. Finally, our approach is validated on the ultralarge size virtual screening data set, retrieving 70% of the top-0.05% of ligands after screening only 2% of the library. Altogether, this work provides a computationally accessible approach for accelerated virtual screening that can serve as a blueprint for the future design of low-compute agents for exploration of the chemical space via large-scale accelerated docking. With recent breakthroughs in protein structure prediction, this method can significantly increase accessibility for the academic community and aid in the rapid discovery of high-affinity hit compounds for various targets.

Principal component analysis highlights the influence of temperature, curvature and cholesterol on conformational dynamics of lipids

Membrane lipids are inherently highly dynamic molecules. Currently, it is difficult to probe the structures of individual lipids experimentally at the timescales corresponding to atomic motions, and consequently molecular dynamics simulations are used widely. In our previous work, we have introduced the principal component analysis (PCA) as a convenient framework for comprehensive quantitative description of lipid motions. Here, we present a newly developed open source script, PCAlipids, which automates the analysis and allows us to refine the approach and test its limitations. We use PCAlipids to determine the influence of temperature, cholesterol and curvature on individual lipids, and show that the most prominent lipid tail scissoring motion is strongly affected by these factors and allows tracking of phase transition. Addition of cholesterol affects the conformations and selectively changes the dynamics of lipid molecules, impacting the large-amplitude motions. Introduction of curvature biases the conformational ensembles towards more extended structures. We hope that the developed approach will be useful for understanding the molecular basis of different processes occurring in lipid membrane systems and will stimulate development of complementary experimental techniques probing the conformations of individual lipid molecules.