RNA3DCNN: Local and global quality assessments of RNA 3D structures using 3D deep convolutional neural networks.
RNA3DCNN: Local and global quality assessments of RNA 3D structures using 3D deep convolutional neural networks.
复制标题
RNA3DCNN：使用 3D 深度卷积神经网络对 RNA 3D 结构进行局部和全局质量评估
DOI：
10.1371/journal.pcbi.1006514
复制
发表时间：
2018-11
影响因子：
4.3
通讯作者：
Wang W
中科院分区：
文献类型：
--
作者：
Li J;Zhu W;Wang J;Li W;Gong S;Zhang J;Wang W
质量评估是RNA三级结构计算预测和设计的关键。迄今为止，已经提出了几种基于知识的统计势，并被证明是有效的，在确定天然和近天然RNA结构。所有这些电位都是基于逆玻尔兹曼公式，而在几何描述符，参考状态和训练数据集的选择不同。通过一种完全不同于传统统计潜力的方法，我们的工作探索了基于3D卷积神经网络（CNN）的方法作为RNA 3D结构质量评估器的能力，该方法使用结构的3D网格表示作为输入，而无需手动提取特征。通过检查每个核苷酸来评估RNA结构，因此我们的方法也可以提供局部质量评估。建立了两组训练样本。第一个包括由高温分子动力学（MD）模拟生成的100万个样本，第二个包括由蒙特卡罗（MC）结构预测生成的100万个样本。对414个RNA的非冗余组进行MD和MC程序。对于两个训练数据集（一个只包括MD训练样本，另一个包括MD和MC训练样本），我们训练了两个神经网络，分别命名为RNA3DCNN_MD和RNA3DCNN_MDMC。前者适用于评估近原生结构，而后者适用于评估覆盖大结构空间的结构。我们测试了我们的方法的性能，并与其他四个传统的评分功能进行了比较。在三个测试数据集中的两个上，我们的方法与最先进的传统评分函数相似，而在第三个测试数据集上，我们的方法远远优于其他评分函数。我们的方法可以从www.example.com下载。
Quality assessment is essential for the computational prediction and design of RNA tertiary structures. To date, several knowledge-based statistical potentials have been proposed and proved to be effective in identifying native and near-native RNA structures. All these potentials are based on the inverse Boltzmann formula, while differing in the choice of the geometrical descriptor, reference state, and training dataset. Via an approach that diverges completely from the conventional statistical potentials, our work explored the power of a 3D convolutional neural network (CNN)-based approach as a quality evaluator for RNA 3D structures, which used a 3D grid representation of the structure as input without extracting features manually. The RNA structures were evaluated by examining each nucleotide, so our method can also provide local quality assessment. Two sets of training samples were built. The first one included 1 million samples generated by high-temperature molecular dynamics (MD) simulations and the second one included 1 million samples generated by Monte Carlo (MC) structure prediction. Both MD and MC procedures were performed for a non-redundant set of 414 RNAs. For two training datasets (one including only MD training samples and the other including both MD and MC training samples), we trained two neural networks, named RNA3DCNN_MD and RNA3DCNN_MDMC, respectively. The former is suitable for assessing near-native structures, while the latter is suitable for assessing structures covering large structural space. We tested the performance of our method and made comparisons with four other traditional scoring functions. On two of three test datasets, our method performed similarly to the state-of-the-art traditional scoring function, and on the third test dataset, our method was far superior to other scoring functions. Our method can be downloaded from https://github.com/lijunRNA/RNA3DCNN.
登录
查看更多内容
DOI：
10.1038/nrc.2017.99
发表时间：
2018-01
期刊：
Nature reviews. Cancer
影响因子：
--
作者：
Anastasiadou E;Jacob LS;Slack FJ
通讯作者：
Slack FJ
影响因子：
4.5
作者：
Flores, Samuel Coulbourn;Altman, Russ B.
通讯作者：
Altman, Russ B.
影响因子：
56.9
作者：
Carleo, Giuseppe;Troyer, Matthias
通讯作者：
Troyer, Matthias
影响因子：
4.5
作者：
Jonikas, Magdalena A.;Radmer, Randall J.;Altman, Russ B.
通讯作者：
Altman, Russ B.
影响因子：
5.8
作者：
Jossinet, F;Westhof, E
通讯作者：
Westhof, E