DNCON2: improved protein contact prediction using two-level deep convolutional neural networks.
DNCON2: improved protein contact prediction using two-level deep convolutional neural networks.
复制标题
DOI：
10.1093/bioinformatics/btx781
复制
发表时间：
2018-05-01
期刊：
影响因子：
--
通讯作者：
Cheng J
中科院分区：
文献类型：
--
作者：
Adhikari B;Hou J;Cheng J
近年来，在预测蛋白质残基-残基接触方面取得了重大进展。正如最近的CASP实验所证明的那样，使用各种基于共同进化的方法和机器学习方法预测的这些接触是从头计算蛋白质结构预测最新进展的关键贡献者。继续发展可靠预测接触图的新方法对于进一步改进从头算结构预测至关重要。本文讨论了一种改进的基于两级深度卷积神经网络的蛋白质接触图谱预测器DNCON2。它由六个卷积神经网络组成——前五个预测6、7.5、8、8.5和10 Å距离阈值的接触，最后一个使用这五个预测作为额外的特征来预测最终的接触图。在CASP10、11和12个实验的自由建模数据集上，DNCON2的平均精度分别为35.0%、50%和53.4%，高于MetaPSICOV在CASP10数据集上的30.6%、MetaPSICOV在CASP11数据集上的34%和Raptor-X在CASP12数据集上的46.3%。我们将DNCON2性能的提高归功于将中短期接触纳入训练、两级预测方法、使用最先进的优化和激活函数，以及一种新颖的深度学习架构，该架构允许卷积层中的每个过滤器访问任意长度蛋白质的所有输入特征。DNCON2的web服务器位于http://sysbio.rnet.missouri.edu/dncon2/，可以下载训练和测试数据集以及CASP10, 11和12的预测免费建模数据集。其源代码可从https://github.com/multicom-toolbox/DNCON2/获得。补充数据可在生物信息学网站获得。
Significant improvements in the prediction of protein residue–residue contacts are observed in the recent years. These contacts, predicted using a variety of coevolution-based and machine learning methods, are the key contributors to the recent progress in ab initio protein structure prediction, as demonstrated in the recent CASP experiments. Continuing the development of new methods to reliably predict contact maps is essential to further improve ab initio structure prediction. In this paper we discuss DNCON2, an improved protein contact map predictor based on two-level deep convolutional neural networks. It consists of six convolutional neural networks—the first five predict contacts at 6, 7.5, 8, 8.5 and 10 Å distance thresholds, and the last one uses these five predictions as additional features to predict final contact maps. On the free-modeling datasets in CASP10, 11 and 12 experiments, DNCON2 achieves mean precisions of 35, 50 and 53.4%, respectively, higher than 30.6% by MetaPSICOV on CASP10 dataset, 34% by MetaPSICOV on CASP11 dataset and 46.3% by Raptor-X on CASP12 dataset, when top L/5 long-range contacts are evaluated. We attribute the improved performance of DNCON2 to the inclusion of short- and medium-range contacts into training, two-level approach to prediction, use of the state-of-the-art optimization and activation functions, and a novel deep learning architecture that allows each filter in a convolutional layer to access all the input features of a protein of arbitrary length. The web server of DNCON2 is at http://sysbio.rnet.missouri.edu/dncon2/ where training and testing datasets as well as the predictions for CASP10, 11 and 12 free-modeling datasets can also be downloaded. Its source code is available at https://github.com/multicom-toolbox/DNCON2/. Supplementary data are available at Bioinformatics online.
登录
查看更多内容
影响因子：
14.9
作者：
Cheng, J;Randall, AZ;Sweredoski, MJ;Baldi, P
通讯作者：
Baldi, P
DOI：
10.1093/bioinformatics/btu458
发表时间：
2014-09-01
期刊：
Bioinformatics (Oxford, England)
影响因子：
--
作者：
Michel M;Hayat S;Skwark MJ;Sander C;Marks DS;Elofsson A
通讯作者：
Elofsson A
影响因子：
3
作者：
Eickholt J;Cheng J
通讯作者：
Cheng J
影响因子：
2.9
作者：
Monastyrskyy B;D'Andrea D;Fidelis K;Tramontano A;Kryshtafovych A
通讯作者：
Kryshtafovych A
影响因子：
2.9
作者：
Monastyrskyy, Bohdan;Fidelis, Krzysztof;Tramontano, Anna;Kryshtafovych, Andriy
通讯作者：
Kryshtafovych, Andriy