Learning protein structure contact maps using neural networks.

机译：使用神经网络学习蛋白质结构接触图。

获取原文

获取原文并翻译 | 示例

页面导航

摘要
著录项
相似文献
相关主题

摘要

Contact maps of proteins are predicted using a neural network-based method. The encoding for the inputs is based on the I-site library developed by Bystroff and others. As a comparison to the I-sites inputs, five other encodings, based on amino acid features, are also used. The five features used are: (1) the polarity of the amino acids, (2) the hydrophobicity of the amino acids, (3) the secondary structure of the amino acid, (4) the amino acid sequence, and (5) all four put together. The set of proteins used comes from the Protein Data Bank and contains 523 nonhomogeneous proteins, which are divided into three different sets and tested. Each of the programs uses 26 separate back-propagation neural networks for training, cross-validating, and testing. The results show that the I-site data used this way did not produce more accurate results than other previous methods such as those of Fariselli and others.

机译：使用基于神经网络的方法预测蛋白质的接触图。输入的编码基于Bystroff等开发的I-site库。为了与I位输入进行比较，还使用了其他五种基于氨基酸特征的编码。使用的五个特征是：（1）氨基酸的极性；（2）氨基酸的疏水性；（3）氨基酸的二级结构；（4）氨基酸序列；以及（5）所有四个放在一起。使用的蛋白质组来自蛋白质数据库，包含523个非均质蛋白质，将其分为三组并进行了测试。每个程序都使用26个单独的反向传播神经网络进行训练，交叉验证和测试。结果表明，与其他以前的方法（例如Fariselli和其他方法）相比，以这种方式使用的I站点数据没有产生更准确的结果。

著录项

作者
Poulson, Vaughn.;
展开▼
作者单位

Utah State University.;

展开▼
授予单位 Utah State University.;
学科 Computer science.
学位 M.S.
年度 2007
页码 39 p.
总页数 39
原文格式 PDF
正文语种 eng
中图分类
关键词

相似文献

外文文献
中文文献
专利

1. Prediction of helix-helix contacts and interacting helices in polytopic membrane proteins using neural networks. [J] . Fuchs A, Kirschner A, Frishman D Proteins: Structure, Function, and Genetics . 2009,第4期

机译：使用神经网络预测多螺旋膜蛋白中的螺旋-螺旋接触和相互作用的螺旋。
2. Improving prediction of protein secondary structure, backbone angles, solvent accessibility and contact numbers by using predicted contact maps and an ensemble of recurrent and residual convolutional neural networks [J] . Hanson Jack, Paliwal Kuldip, Litfin Thomas, Bioinformatics . 2019,第14期

机译：通过使用预测的联系地图和经常性和残余卷积神经网络的集合来改善蛋白质二级结构，骨干角，溶剂可访问性和接触号的预测
3. Detecting distant-homology protein structures by aligning deep neural-network based contact maps [J] . Wei Zheng, Qiqige Wuyun, Yang Li, PLoS Computational Biology . 2019,第10期

机译：通过对齐基于深度神经网络的接触图来检测远距同源蛋白质结构
4. Prediction of Contact Maps in Proteins Based on Recurrent Neural Network with Bias Units [C] . Guixia Liu, Chunguang Zhou, Yuanxian Zhu, International Symposium on Neural Networks(ISNN 2005) pt.3; 20050530-0601; Chongqing(CN) . 2005

机译：基于带偏差单元的递归神经网络的蛋白质接触图预测
5. Deciphering Protein Complex Structures from Cryo-Electron Microscopy Maps Using Deep Learning [D] . Pfab, Jonas. 2020

机译：使用深度学习从冷冻电子显微镜地图中解密蛋白质复合结构
6. Detecting distant-homology protein structures by aligning deep neural-network based contact maps [O] . Wei Zheng, Qiqige Wuyun, Yang Li, 2019

机译：通过对齐基于深度神经网络的接触图来检测远距同源蛋白质结构
7. CNNcon: improved protein contact maps prediction using cascaded neural networks. [O] . Wang Ding, Jiang Xie, Dongbo Dai, 2013

机译：CNNcon：使用级联神经网络改进蛋白质接触图预测。

Learning protein structure contact maps using neural networks.

摘要

著录项

相似文献

相关主题

期刊订阅