Learning Deep Architectures for Interaction Prediction in Structure-based Virtual Screening
This work addresses the need for improved virtual screening in drug discovery by proposing a novel deep learning approach and a more robust benchmark, though it appears incremental as it builds on existing methods and datasets.
The authors tackled the problem of predicting protein-ligand interactions in virtual screening by introducing a deep learning architecture that generates fixed-sized fingerprints for proteins and small molecules, using learnable atom convolution and softmax operations, and predicting binding potential through inner-product calculations. They also created a new benchmark dataset from DUD-E and PDBBind databases, addressing limitations in existing benchmarks for machine learning methods.
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.