Bionoi: A Voronoi Diagram-Based Representation of Ligand-Binding Sites in Proteins for Machine Learning Applications.

@article{Feinstein2021BionoiAV,
  title={Bionoi: A Voronoi Diagram-Based Representation of Ligand-Binding Sites in Proteins for Machine Learning Applications.},
  author={Joseph Feinstein and Wentao Shi and J. Ramanujam and Michal Brylinski},
  journal={Methods in molecular biology},
  year={2021},
  volume={2266},
  pages={
          299-312
        }
}
Bionoi is a new software to generate Voronoi representations of ligand-binding sites in proteins for machine learning applications. Unlike many other deep learning models in biomedicine, Bionoi utilizes off-the-shelf convolutional neural network architectures, reducing the development work without sacrificing the performance. When initially generating images of binding sites, users have the option to color the Voronoi cells based on either one of six structural, physicochemical, and… 
Graphsite: Ligand-binding site classification using Deep Graph Neural Network
TLDR
A state-of-the-art GNN model is trained to capture the intrinsic characteristics of these binding sites and classify them and achieves test accuracy of 81.28% on classifying 14 binding site classes.

References

SHOWING 1-10 OF 15 REFERENCES
BionoiNet: ligand-binding site classification with off-the-shelf deep neural network
TLDR
This work developed BionoiNet, a new deep learning-based framework implementing a popular ResNet model for image classification, which transforms the molecular structures of ligand-binding sites to 2D Voronoi diagrams, which are then used as the input to a pretrained convolutional neural network classifier.
DeepSite: protein‐binding site predictor using 3D‐convolutional neural networks
TLDR
This work presents a novel knowledge‐based approach that uses state‐of‐the‐art convolutional neural networks, where the algorithm is learned by examples, and demonstrates superior performance to two other competitive algorithmic strategies.
DeepDrug3D: Classification of ligand-binding pockets in proteins with a convolutional neural network
TLDR
DeepDrug3D is presented, a new approach to characterize and classify binding pockets in proteins with deep learning that employs a state-of-the-art convolutional neural network in which biomolecular structures are represented as voxels assigned interaction energy-based attributes.
LigVoxel: inpainting binding pockets using 3D‐convolutional neural networks
TLDR
It is shown that an end‐to‐end deep learning framework trained on experimental protein‐ligand complexes can generate spatial images of ligand chemical properties like occupancy, aromaticity and donor‐acceptor matching the protein pocket.
Predicting protein-ligand binding residues with deep convolutional neural networks
TLDR
A new sequence-based approach called DeepCSeqSite for ab initio protein-ligand binding residue prediction that significantly outperforms existing sequence- based and 3D-structure-based methods, including COACH.
Mol2vec: Unsupervised Machine Learning Approach with Chemical Intuition
TLDR
Mol2vec can be easily combined with ProtVec, which employs the same Word2vec concept on protein sequences, resulting in a proteochemometric approach that is alignment-independent and thus can also be easily used for proteins with low sequence similarities.
DeepSF: deep convolutional neural network for mapping protein sequences to folds
TLDR
A deep 1D‐convolution neural network (DeepSF) is developed to directly classify any protein sequence into one of 1195 known folds, which is useful for both fold recognition and the study of sequence‐structure relationship.
POPS: a fast algorithm for solvent accessible surface areas at atomic and residue level
TLDR
The residue based approach POPS-R, has been devised as a useful tool for the analysis of large macromolecular assemblies like the ribosome and it is specially suited for the refinement of low resolution structures.
ProtVec: A Continuous Distributed Representation of Biological Sequences
TLDR
By only providing sequence data for various proteins into this model, information about protein structure can be determined with high accuracy, and this so-called embedding model needs to be trained only once and can be used to ascertain a diverse set of information regarding the proteins of interest.
Build-Up Algorithm for Atomic Correspondence between Chemical Structures
  • T. Kawabata
  • Computer Science, Chemistry
    J. Chem. Inf. Model.
  • 2011
TLDR
A new program, called "kcombu" is developed using a build-up algorithm, which is a type of the greedy heuristic algorithms, which shows that the TD-MCS has the best agreement value among the other types of MCS.
...
...