Voxel-based convolutional neural networks have been used to predict protein-ligand binding affinity and classify protein binding pockets. However, memory usage of voxel-based approaches increases cubically with the spatial resolution. It is hence necessary and beneficial to develop memory-efficient methods that enable the voxel representation of the protein-ligand interaction in a much higher resolution. In this study, we implement an efficient hierarchical octree data structure that performs a tree search to only split the voxels where protein or ligand atoms is present into smaller octants while keeping non-occupied space as big octants. In this way, a very high resolution can be achieve around the protein-ligand interaction surfaces without significantly blowing up the number of voxels in other regions. Although hierarchical octree data structure has been widely used in 3D graphics and 3D game engines to improve memory efficiency for 3D modeling, it has not been applied and validated in molecular representation. We carry out a series of experiments to demonstrate the substantial advantages of the octree surface representation with respect to memory efficiency and visualization accuracy. We also implement an adaptive-resolution octree-based volumetric representation of protein-ligand complexes. Unlike classical volumetric- or octree-based methods that represent three-dimensional objects with single-size voxels, our adaptive implementation supports multi-resolution characterization of different regions within protein-ligand complexes.

OctSurf: Efficient hierarchical voxel-based representation for protein-ligand interaction surfaces

Macromolecular Chemistry: The Second Century

[Author Benefits](https://scimeetings.acs.org/?utm_source=pubs_content_marketing&utm_medium=website&utm_campaign=0320_MCF_NPI_Launch_Spring_Homepage&ref=pubs_content_marketing)

[How to use SciMeetings](https://storage.googleapis.com/dl.morressier.com/acs/SciMeetings-ACSSpring2020-HowToGuide_M3020-Final.pdf)

[Frequently Asked Questions](https://storage.googleapis.com/dl.morressier.com/acs/SciMtng_ACSNMSp2020_FAQ_M3120_Final.pdf)

ACS Spring 2020 National Meeting & Expo

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.