Learning Deep Architectures for Interaction Prediction in Structure-based Virtual Screening

TL;DR

This paper presents a novel deep learning architecture for predicting protein-ligand interactions in virtual screening, introducing learnable atom convolution, a new benchmark dataset, and demonstrating improved predictive capabilities.

Contribution

The work introduces a new deep learning model with learnable atom convolution for virtual screening and provides a new benchmark dataset for more effective evaluation.

Findings

Deep learning fingerprints improve interaction prediction.

The new benchmark dataset offers a more rigorous testing environment.

Traditional datasets may be insufficient for ML-based virtual screening.

Abstract

We introduce a deep learning architecture for structure-based virtual screening that generates fixed-sized fingerprints of proteins and small molecules by applying learnable atom convolution and softmax operations to each compound separately. These fingerprints are further transformed non-linearly, their inner-product is calculated and used to predict the binding potential. Moreover, we show that widely used benchmark datasets may be insufficient for testing structure-based virtual screening methods that utilize machine learning. Therefore, we introduce a new benchmark dataset, which we constructed based on DUD-E and PDBBind databases.