Pseudouridine (Ψ) is the first discovered and the most prevalent posttranscriptional modification, which has been widely studied during the past decades. Pseudouridine was observed in almost all kinds of RNAs and shown to have important biological functions. Currently, the time-consuming and high-cost procedures of experimental approaches limit its uses in real-life Ψ site detection. Alternatively, by taking advantage of the explosive growth of Ψ sequencing data, the computational methods may provide a more cost-effective avenue. To date, the existing mouse Ψ site predictors were all developed based on sequence-derived features, and their performance can be further improved by adding the domain knowledge derived feature. Therefore, it is highly desirable to propose a genomic feature-based computational method to increase the accuracy and efficiency of the identification of Ψ RNA modification in the mouse transcriptome. In our study, a predictive framework PSI-MOUSE was built. Besides the conventional sequence-based features, PSI-MOUSE first introduced 38 additional genomic features derived from the mouse genome, which achieved a satisfactory improvement in the prediction performance, compared with other existing models. Moreover, PSI-MOUSE also features in automatically annotating the putative Ψ sites with diverse types of posttranscriptional regulations (RNA-binding protein [RBP]-binding regions, miRNA-RNA interactions, and splicing sites), which can serve as a useful research tool for the study of Ψ RNA modification in the mouse genome. Finally, 3282 experimentally validated mouse Ψ sites were also collected in a database with customized query functions. For the convenience of academic users, a website was built to provide a user-friendly interface for the query and analysis on the database. The website is freely accessible at www.xjtlu.edu.cn/biologicalsciences/psimouse and http://psimouse.rnamd.com. We introduced the genome-derived features to mouse for the first time, and we achieved a good performance in mouse Ψ site prediction. Compared with the existing state-of-art methods, our newly developed approach PSI-MOUSE obtained a substantial improvement in prediction accuracy, marking the reliable contributions of genomic features for the prediction of RNA modifications in a species other than human.

假性尿苷（se）是最早发现的，也是最普遍的转录后修饰，在过去的几十年中进行了广泛的研究。几乎在所有种类的RNA中都观察到了假性尿苷，并显示了其重要的生物学功能。目前，实验方法耗时且成本高昂的程序限制了其在现实生活中的站点检测中的使用。或者，通过利用Ψ测序数据的爆炸性增长，计算方法可以提供更具成本效益的途径。迄今为止，现有的小鼠Ψ位点预测子都是基于序列衍生的特征开发的，通过添加领域知识衍生的特征可以进一步提高其性能。因此，迫切需要提出一种基于基因组特征的计算方法，以提高在小鼠转录组中鉴定RNA修饰的准确性和效率。在我们的研究中，建立了一个预测框架PSI-MOUSE。除了传统的基于序列的特征外，PSI-MOUSE还首次引入了38个源自小鼠基因组的其他基因组特征，与其他现有模型相比，它们在预测性能方面取得了令人满意的提高。此外，PSI-MOUSE还具有通过各种类型的转录后调控（RNA结合蛋白[RBP]结合区，miRNA-RNA相互作用和剪接位点）自动注释可能的Ψ位点的功能，可以用作有用的研究工具用于研究小鼠基因组中的RNA修饰。最后，还通过具有自定义查询功能的数据库收集了3282个经过实验验证的Mouse站点。为了方便学术用户，建立了一个网站，以提供用户友好的界面来查询和分析数据库。该网站可从www.xjtlu.edu.cn/biologicalsciences/psimouse和http://psimouse.rnamd.com免费访问。我们首次将基因组衍生功能引入小鼠，并在小鼠Ψ位点预测中取得了良好的性能。与现有的最新技术相比，我们新开发的方法PSI-MOUSE在预测准确性上有了实质性的提高，标志着基因组特征对人类以外物种的RNA修饰预测的可靠贡献。为了方便学术用户，建立了一个网站，以提供用户友好的界面来查询和分析数据库。该网站可从www.xjtlu.edu.cn/biologicalsciences/psimouse和http://psimouse.rnamd.com免费访问。我们首次将基因组衍生功能引入小鼠，并在小鼠Ψ位点预测中取得了良好的性能。与现有的最新方法相比，我们新开发的PSI-MOUSE方法在预测准确性上有了实质性的提高，标志着基因组特征对人类以外物种的RNA修饰预测的可靠贡献。为了方便学术用户，建立了一个网站，以提供用户友好的界面来查询和分析数据库。该网站可从www.xjtlu.edu.cn/biologicalsciences/psimouse和http://psimouse.rnamd.com免费访问。我们首次将基因组衍生功能引入小鼠，并在小鼠Ψ位点预测中取得了良好的性能。与现有的最新技术相比，我们新开发的方法PSI-MOUSE在预测准确性上有了实质性的提高，标志着基因组特征对人类以外物种的RNA修饰预测的可靠贡献。cn / biologicalsciences / psimouse和http://psimouse.rnamd.com。我们首次将基因组衍生功能引入小鼠，并在小鼠Ψ位点预测中取得了良好的性能。与现有的最新技术相比，我们新开发的方法PSI-MOUSE在预测准确性上有了实质性的提高，标志着基因组特征对人类以外物种的RNA修饰预测的可靠贡献。cn / biologicalsciences / psimouse和http://psimouse.rnamd.com。我们首次将基因组衍生功能引入小鼠，并在小鼠Ψ位点预测中取得了良好的性能。与现有的最新技术相比，我们新开发的方法PSI-MOUSE在预测准确性上有了实质性的提高，标志着基因组特征对人类以外物种的RNA修饰预测的可靠贡献。