AK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks
Accurate prediction of the binding affinity of a protein-ligand complex is essential for efficient and successful rational drug design. Therefore, many binding affinity prediction methods have been developed. In recent years, since deep learning technology has become powerful, it is also implemented...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2020-11-01
|
Series: | International Journal of Molecular Sciences |
Subjects: | |
Online Access: | https://www.mdpi.com/1422-0067/21/22/8424 |
_version_ | 1797548414908497920 |
---|---|
author | Yongbeom Kwon Woong-Hee Shin Junsu Ko Juyong Lee |
author_facet | Yongbeom Kwon Woong-Hee Shin Junsu Ko Juyong Lee |
author_sort | Yongbeom Kwon |
collection | DOAJ |
description | Accurate prediction of the binding affinity of a protein-ligand complex is essential for efficient and successful rational drug design. Therefore, many binding affinity prediction methods have been developed. In recent years, since deep learning technology has become powerful, it is also implemented to predict affinity. In this work, a new neural network model that predicts the binding affinity of a protein-ligand complex structure is developed. Our model predicts the binding affinity of a complex using the ensemble of multiple independently trained networks that consist of multiple channels of 3-D convolutional neural network layers. Our model was trained using the 3772 protein-ligand complexes from the refined set of the PDBbind-2016 database and tested using the core set of 285 complexes. The benchmark results show that the Pearson correlation coefficient between the predicted binding affinities by our model and the experimental data is 0.827, which is higher than the state-of-the-art binding affinity prediction scoring functions. Additionally, our method ranks the relative binding affinities of possible multiple binders of a protein quite accurately, comparable to the other scoring functions. Last, we measured which structural information is critical for predicting binding affinity and found that the complementarity between the protein and ligand is most important. |
first_indexed | 2024-03-10T14:59:02Z |
format | Article |
id | doaj.art-928081b5365343caa6a3f65453cb56a3 |
institution | Directory Open Access Journal |
issn | 1661-6596 1422-0067 |
language | English |
last_indexed | 2024-03-10T14:59:02Z |
publishDate | 2020-11-01 |
publisher | MDPI AG |
record_format | Article |
series | International Journal of Molecular Sciences |
spelling | doaj.art-928081b5365343caa6a3f65453cb56a32023-11-20T20:23:10ZengMDPI AGInternational Journal of Molecular Sciences1661-65961422-00672020-11-012122842410.3390/ijms21228424AK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural NetworksYongbeom Kwon0Woong-Hee Shin1Junsu Ko2Juyong Lee3Department of Chemistry, Kangwon National University, Gangwon-do, Chuncheon 24341, KoreaDepartment of Chemical Science Education, Sunchon National University, Jeollanam-do, Suncheon 57922, KoreaArontier, 241 Gangnam-daero, Seocho-gu, Seoul 06735, KoreaDepartment of Chemistry, Kangwon National University, Gangwon-do, Chuncheon 24341, KoreaAccurate prediction of the binding affinity of a protein-ligand complex is essential for efficient and successful rational drug design. Therefore, many binding affinity prediction methods have been developed. In recent years, since deep learning technology has become powerful, it is also implemented to predict affinity. In this work, a new neural network model that predicts the binding affinity of a protein-ligand complex structure is developed. Our model predicts the binding affinity of a complex using the ensemble of multiple independently trained networks that consist of multiple channels of 3-D convolutional neural network layers. Our model was trained using the 3772 protein-ligand complexes from the refined set of the PDBbind-2016 database and tested using the core set of 285 complexes. The benchmark results show that the Pearson correlation coefficient between the predicted binding affinities by our model and the experimental data is 0.827, which is higher than the state-of-the-art binding affinity prediction scoring functions. Additionally, our method ranks the relative binding affinities of possible multiple binders of a protein quite accurately, comparable to the other scoring functions. Last, we measured which structural information is critical for predicting binding affinity and found that the complementarity between the protein and ligand is most important.https://www.mdpi.com/1422-0067/21/22/8424protein-ligand binding affinityconvolutional neural networkResNextdeep learningbinding affinity predictiondocking score |
spellingShingle | Yongbeom Kwon Woong-Hee Shin Junsu Ko Juyong Lee AK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks International Journal of Molecular Sciences protein-ligand binding affinity convolutional neural network ResNext deep learning binding affinity prediction docking score |
title | AK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks |
title_full | AK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks |
title_fullStr | AK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks |
title_full_unstemmed | AK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks |
title_short | AK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks |
title_sort | ak score accurate protein ligand binding affinity prediction using an ensemble of 3d convolutional neural networks |
topic | protein-ligand binding affinity convolutional neural network ResNext deep learning binding affinity prediction docking score |
url | https://www.mdpi.com/1422-0067/21/22/8424 |
work_keys_str_mv | AT yongbeomkwon akscoreaccurateproteinligandbindingaffinitypredictionusinganensembleof3dconvolutionalneuralnetworks AT woongheeshin akscoreaccurateproteinligandbindingaffinitypredictionusinganensembleof3dconvolutionalneuralnetworks AT junsuko akscoreaccurateproteinligandbindingaffinitypredictionusinganensembleof3dconvolutionalneuralnetworks AT juyonglee akscoreaccurateproteinligandbindingaffinitypredictionusinganensembleof3dconvolutionalneuralnetworks |