基于双层BiGRU网络的哺乳动物组织m6A甲基化位点预测

目的 N⁶-甲基化腺苷（N⁶-methyladenosine,m⁶A）是RNA中最常见、最丰富的化学修饰,在很多生物过程中发挥着重要作用。目前已经发展了一些预测m⁶A甲基化位点的计算方法。然而,这些方法在针对不同物种或不同组织时,缺乏稳健性。为了提升对不同组织中m⁶A甲基化位点预测的稳健性,本文提出一种能结合序列反向信息来提取数据更高级特征的双层双向门控循环单元（bidirectional gated recurrent unit,BiGRU）网络模型。方法 本文选取具有代表性的哺乳动物组织m⁶A甲基化位点数据集作为训练数据,通过对模型网络、网络结构、层数和优化器等进行搭配,构建双层BiGRU网络。结果 将模型应用于人类、小鼠和大鼠共11个组织的m⁶A甲基化位点预测上,并与其他方法在这11个组织上的预测能力进行了全面的比较。结果表明,本文构建的模型平均预测接受者操作特征曲线下面积（area under the receiver operating characteristic curve,AUC）达到93.72%,与目前最好的预测方法持平,而预测准确率（accuracy,ACC）、敏感性（sensitivity,SN）、特异性（specificity,SP）和马修斯相关系数（Matthews correlation coefficient,MCC）分别为90.07%、90.30%、89.84%和80.17%,均高于目前的m⁶A甲基化位点预测方法。结论 和已有研究方法相比,本文方法对11个哺乳动物组织的m⁶A甲基化位点的预测准确性均达到最高,说明本文方法具有较好的泛化能力。

Prediction of m6A Methylation Sites in Mammalian Tissues Based on a Double-layer BiGRU Network

Objective N⁶-methyladenosine (m⁶A) is the most common and abundant chemical modification in RNA and plays an important role in many biological processes. Several computational methods have been developed to predict m⁶A methylation sites. However, these methods lack robustness when targeting different species or different tissues. To improve the robustness of the prediction performance of m⁶A methylation sites in different tissues, this paper proposed a double-layer bidirectional gated recurrent unit (BiGRU) network model that combines reverse sequence information to extract higher-level features of the data.Methods Some representative mammalian tissue m⁶A methylation site datasets were selected as the training datasets. Based on a BiGRU, a double-layer BiGRU network was constructed by collocation of the model network, the model structure, the number of layers and the optimizer.Results The model was applied to predict m⁶A methylation sites in 11 human, mouse and rat tissues, and the prediction performance was compared with that of other methods using the same tissues. The results demonstrated that the average area under the receiver operating characteristic curve (AUC) predicted by the proposed model reached 93.72%, equaling that of the best prediction method at present. The values of accuracy (ACC), sensitivity (SN), specificity (SP) and Matthews correlation coefficient (MCC) were 90.07%, 90.30%, 89.84% and 80.17%, respectively, which were higher than those of the current methods for predicting m⁶A methylation sites.Conclusion Compared with that of existing research methods, the prediction accuracy of the double-layer BiGRU network was the highest for identifying m⁶A methylation sites in the 11 tissues, indicating that the method proposed in this study has an excellent generalizability.