ROCS-Derived Features for Virtual Screening

TL;DR

This paper introduces new color similarity features derived from ROCS that, when weighted by machine learning, significantly enhance virtual screening performance over standard ROCS methods.

Contribution

The study decomposes ROCS color features into novel components and demonstrates their effectiveness in improving virtual screening accuracy through machine learning.

Findings

Significant improvement in ROC AUC scores with new features

Color component features enhance virtual screening performance

Machine learning effectively weights color features

Abstract

Rapid overlay of chemical structures (ROCS) is a standard tool for the calculation of 3D shape and chemical ("color") similarity. ROCS uses unweighted sums to combine many aspects of similarity, yielding parameter-free models for virtual screening. In this report, we decompose the ROCS color force field into "color components" and "color atom overlaps", novel color similarity features that can be weighted in a system-specific manner by machine learning algorithms. In cross-validation experiments, these additional features significantly improve virtual screening performance (ROC AUC scores) relative to standard ROCS.