Improvement in positional accuracy of neural-network predicted hydration sites of proteins by incorporating atomic details of water-protein interactions and site-searching algorithm

Visualization of hydration structures over the entire protein surface is necessary to understand why the aqueous environment is essential for protein folding and functions. However, it is still difficult for experiments. Recently, we developed a convolutional neural network (CNN) to predict the prob...

Full description

Saved in:
Bibliographic Details
Main Authors: Kochi Sato, Masayoshi Nakasako
Format: Article
Language:English
Published: The Biophysical Society of Japan 2025-03-01
Series:Biophysics and Physicobiology
Subjects:
Online Access:https://doi.org/10.2142/biophysico.bppb-v22.0004
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849705347377266688
author Kochi Sato
Masayoshi Nakasako
author_facet Kochi Sato
Masayoshi Nakasako
author_sort Kochi Sato
collection DOAJ
description Visualization of hydration structures over the entire protein surface is necessary to understand why the aqueous environment is essential for protein folding and functions. However, it is still difficult for experiments. Recently, we developed a convolutional neural network (CNN) to predict the probability distribution of hydration water molecules over protein surfaces and in protein cavities. The deep network was optimized using solely the distribution patterns of protein atoms surrounding each hydration water molecule in high-resolution X-ray crystal structures and successfully provided probability distributions of hydration water molecules. Despite the effectiveness of the probability distribution, the positional differences of the predicted positions obtained from the local maxima as predicted sites remained inadequate in reproducing the hydration sites in the crystal structure models. In this work, we modified the deep network by subdividing atomic classes based on the electronic properties of atoms composing amino acids. In addition, the exclusion volumes of each protein atom and hydration water molecule were taken to predict the hydration sites from the probability distribution. These information on chemical properties of atoms leads to an improvement in positional prediction accuracy. We selected the best CNN from 47 CNNs constructed by systematically varying the number of channels and layers of neural networks. Here, we report the improvements in prediction accuracy by the reorganized CNN together with the details in the architecture, training data, and peak search algorithm.
format Article
id doaj-art-a6d69eea2203457ba9a4f12458caac2e
institution DOAJ
issn 2189-4779
language English
publishDate 2025-03-01
publisher The Biophysical Society of Japan
record_format Article
series Biophysics and Physicobiology
spelling doaj-art-a6d69eea2203457ba9a4f12458caac2e2025-08-20T03:16:29ZengThe Biophysical Society of JapanBiophysics and Physicobiology2189-47792025-03-012210.2142/biophysico.bppb-v22.0004Improvement in positional accuracy of neural-network predicted hydration sites of proteins by incorporating atomic details of water-protein interactions and site-searching algorithmKochi Sato0Masayoshi Nakasako1Department of Physics, Faculty of Science and Technology, Keio University, Yokohama, Kanagawa 223-8522, JapanDepartment of Physics, Faculty of Science and Technology, Keio University, Yokohama, Kanagawa 223-8522, JapanVisualization of hydration structures over the entire protein surface is necessary to understand why the aqueous environment is essential for protein folding and functions. However, it is still difficult for experiments. Recently, we developed a convolutional neural network (CNN) to predict the probability distribution of hydration water molecules over protein surfaces and in protein cavities. The deep network was optimized using solely the distribution patterns of protein atoms surrounding each hydration water molecule in high-resolution X-ray crystal structures and successfully provided probability distributions of hydration water molecules. Despite the effectiveness of the probability distribution, the positional differences of the predicted positions obtained from the local maxima as predicted sites remained inadequate in reproducing the hydration sites in the crystal structure models. In this work, we modified the deep network by subdividing atomic classes based on the electronic properties of atoms composing amino acids. In addition, the exclusion volumes of each protein atom and hydration water molecule were taken to predict the hydration sites from the probability distribution. These information on chemical properties of atoms leads to an improvement in positional prediction accuracy. We selected the best CNN from 47 CNNs constructed by systematically varying the number of channels and layers of neural networks. Here, we report the improvements in prediction accuracy by the reorganized CNN together with the details in the architecture, training data, and peak search algorithm.https://doi.org/10.2142/biophysico.bppb-v22.0004deep learningprotein hydrationhydrophobic hydrationx-ray crystallographyartificial intelligence
spellingShingle Kochi Sato
Masayoshi Nakasako
Improvement in positional accuracy of neural-network predicted hydration sites of proteins by incorporating atomic details of water-protein interactions and site-searching algorithm
Biophysics and Physicobiology
deep learning
protein hydration
hydrophobic hydration
x-ray crystallography
artificial intelligence
title Improvement in positional accuracy of neural-network predicted hydration sites of proteins by incorporating atomic details of water-protein interactions and site-searching algorithm
title_full Improvement in positional accuracy of neural-network predicted hydration sites of proteins by incorporating atomic details of water-protein interactions and site-searching algorithm
title_fullStr Improvement in positional accuracy of neural-network predicted hydration sites of proteins by incorporating atomic details of water-protein interactions and site-searching algorithm
title_full_unstemmed Improvement in positional accuracy of neural-network predicted hydration sites of proteins by incorporating atomic details of water-protein interactions and site-searching algorithm
title_short Improvement in positional accuracy of neural-network predicted hydration sites of proteins by incorporating atomic details of water-protein interactions and site-searching algorithm
title_sort improvement in positional accuracy of neural network predicted hydration sites of proteins by incorporating atomic details of water protein interactions and site searching algorithm
topic deep learning
protein hydration
hydrophobic hydration
x-ray crystallography
artificial intelligence
url https://doi.org/10.2142/biophysico.bppb-v22.0004
work_keys_str_mv AT kochisato improvementinpositionalaccuracyofneuralnetworkpredictedhydrationsitesofproteinsbyincorporatingatomicdetailsofwaterproteininteractionsandsitesearchingalgorithm
AT masayoshinakasako improvementinpositionalaccuracyofneuralnetworkpredictedhydrationsitesofproteinsbyincorporatingatomicdetailsofwaterproteininteractionsandsitesearchingalgorithm