Jared M Sagendorf, Raktim Mitra, Jiawei Huang, Xiaojiang S Chen, Remo Rohs
{"title":"利用图神经网络进行基于结构的蛋白质-核酸结合预测。","authors":"Jared M Sagendorf, Raktim Mitra, Jiawei Huang, Xiaojiang S Chen, Remo Rohs","doi":"10.1007/s12551-024-01201-w","DOIUrl":null,"url":null,"abstract":"<p><p>Protein-nucleic acid (PNA) binding plays critical roles in the transcription, translation, regulation, and three-dimensional organization of the genome. Structural models of proteins bound to nucleic acids (NA) provide insights into the chemical, electrostatic, and geometric properties of the protein structure that give rise to NA binding but are scarce relative to models of unbound proteins. We developed a deep learning approach for predicting PNA binding given the unbound structure of a protein that we call PNAbind. Our method utilizes graph neural networks to encode the spatial distribution of physicochemical and geometric properties of protein structures that are predictive of NA binding. Using global physicochemical encodings, our models predict the overall binding function of a protein, and using local encodings, they predict the location of individual NA binding residues. Our models can discriminate between specificity for DNA or RNA binding, and we show that predictions made on computationally derived protein structures can be used to gain mechanistic understanding of chemical and structural features that determine NA recognition. Binding site predictions were validated against benchmark datasets, achieving AUROC scores in the range of 0.92-0.95. We applied our models to the HIV-1 restriction factor APOBEC3G and showed that our model predictions are consistent with and help explain experimental RNA binding data.</p><p><strong>Supplementary information: </strong>The online version contains supplementary material available at 10.1007/s12551-024-01201-w.</p>","PeriodicalId":9094,"journal":{"name":"Biophysical reviews","volume":null,"pages":null},"PeriodicalIF":4.9000,"publicationDate":"2024-06-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11427629/pdf/","citationCount":"0","resultStr":"{\"title\":\"Structure-based prediction of protein-nucleic acid binding using graph neural networks.\",\"authors\":\"Jared M Sagendorf, Raktim Mitra, Jiawei Huang, Xiaojiang S Chen, Remo Rohs\",\"doi\":\"10.1007/s12551-024-01201-w\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p><p>Protein-nucleic acid (PNA) binding plays critical roles in the transcription, translation, regulation, and three-dimensional organization of the genome. Structural models of proteins bound to nucleic acids (NA) provide insights into the chemical, electrostatic, and geometric properties of the protein structure that give rise to NA binding but are scarce relative to models of unbound proteins. We developed a deep learning approach for predicting PNA binding given the unbound structure of a protein that we call PNAbind. Our method utilizes graph neural networks to encode the spatial distribution of physicochemical and geometric properties of protein structures that are predictive of NA binding. Using global physicochemical encodings, our models predict the overall binding function of a protein, and using local encodings, they predict the location of individual NA binding residues. Our models can discriminate between specificity for DNA or RNA binding, and we show that predictions made on computationally derived protein structures can be used to gain mechanistic understanding of chemical and structural features that determine NA recognition. Binding site predictions were validated against benchmark datasets, achieving AUROC scores in the range of 0.92-0.95. We applied our models to the HIV-1 restriction factor APOBEC3G and showed that our model predictions are consistent with and help explain experimental RNA binding data.</p><p><strong>Supplementary information: </strong>The online version contains supplementary material available at 10.1007/s12551-024-01201-w.</p>\",\"PeriodicalId\":9094,\"journal\":{\"name\":\"Biophysical reviews\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":4.9000,\"publicationDate\":\"2024-06-26\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11427629/pdf/\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Biophysical reviews\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1007/s12551-024-01201-w\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"2024/6/1 0:00:00\",\"PubModel\":\"eCollection\",\"JCR\":\"Q1\",\"JCRName\":\"BIOPHYSICS\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Biophysical reviews","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1007/s12551-024-01201-w","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2024/6/1 0:00:00","PubModel":"eCollection","JCR":"Q1","JCRName":"BIOPHYSICS","Score":null,"Total":0}
引用次数: 0
摘要
蛋白质与核酸(PNA)的结合在基因组的转录、翻译、调控和三维组织中起着至关重要的作用。与核酸(NA)结合的蛋白质结构模型能让人深入了解导致核酸结合的蛋白质结构的化学、静电和几何特性,但相对于未结合的蛋白质模型来说,这种模型还很缺乏。我们开发了一种深度学习方法,用于根据蛋白质的非结合结构预测 PNA 结合,我们称之为 PNAbind。我们的方法利用图神经网络对蛋白质结构的物理化学和几何特性的空间分布进行编码,从而预测 NA 的结合。利用全局理化编码,我们的模型可以预测蛋白质的整体结合功能;利用局部编码,我们的模型可以预测单个 NA 结合残基的位置。我们的模型可以区分 DNA 或 RNA 结合的特异性,我们还展示了通过计算得出的蛋白质结构预测结果可用于从机理上理解决定 NA 识别的化学和结构特征。根据基准数据集对结合位点预测进行了验证,AUROC 得分在 0.92-0.95 之间。我们将模型应用于 HIV-1 限制因子 APOBEC3G,结果表明我们的模型预测与实验 RNA 结合数据一致,并有助于解释这些数据:在线版本包含补充材料,可查阅 10.1007/s12551-024-01201-w。
Structure-based prediction of protein-nucleic acid binding using graph neural networks.
Protein-nucleic acid (PNA) binding plays critical roles in the transcription, translation, regulation, and three-dimensional organization of the genome. Structural models of proteins bound to nucleic acids (NA) provide insights into the chemical, electrostatic, and geometric properties of the protein structure that give rise to NA binding but are scarce relative to models of unbound proteins. We developed a deep learning approach for predicting PNA binding given the unbound structure of a protein that we call PNAbind. Our method utilizes graph neural networks to encode the spatial distribution of physicochemical and geometric properties of protein structures that are predictive of NA binding. Using global physicochemical encodings, our models predict the overall binding function of a protein, and using local encodings, they predict the location of individual NA binding residues. Our models can discriminate between specificity for DNA or RNA binding, and we show that predictions made on computationally derived protein structures can be used to gain mechanistic understanding of chemical and structural features that determine NA recognition. Binding site predictions were validated against benchmark datasets, achieving AUROC scores in the range of 0.92-0.95. We applied our models to the HIV-1 restriction factor APOBEC3G and showed that our model predictions are consistent with and help explain experimental RNA binding data.
Supplementary information: The online version contains supplementary material available at 10.1007/s12551-024-01201-w.
期刊介绍:
Biophysical Reviews aims to publish critical and timely reviews from key figures in the field of biophysics. The bulk of the reviews that are currently published are from invited authors, but the journal is also open for non-solicited reviews. Interested authors are encouraged to discuss the possibility of contributing a review with the Editor-in-Chief prior to submission. Through publishing reviews on biophysics, the editors of the journal hope to illustrate the great power and potential of physical techniques in the biological sciences, they aim to stimulate the discussion and promote further research and would like to educate and enthuse basic researcher scientists and students of biophysics. Biophysical Reviews covers the entire field of biophysics, generally defined as the science of describing and defining biological phenomenon using the concepts and the techniques of physics. This includes but is not limited by such areas as: - Bioinformatics - Biophysical methods and instrumentation - Medical biophysics - Biosystems - Cell biophysics and organization - Macromolecules: dynamics, structures and interactions - Single molecule biophysics - Membrane biophysics, channels and transportation