Recent studies have revealed that N7-methylguanosine(m ⁷ G) plays a pivotal role in various biological processes and disease pathogenesis. To date, transcriptome-wide m ⁷ G modification sites have been identified by high-throughput sequencing approaches, and some related information has been recorded in a few biological databases. However, the mechanism of site action in disease remains uncharted. Wet experiments can help identify true m ⁷ G sites with high confidence, but it is time-consuming to find the true ones in such a large number of sites, which will also cost too much. Thus, computational methods are emergently needed to predict the associations between m ⁷ G sites and various diseases, thus help to uncover potential active sites for specific diseases. In this article, we proposed a bounded robust principal component analysis (BRPCA) method to predict unknown m ⁷ G-disease association based on similarity information. Importantly, BRPCA tolerates the noise and redundancy existing in association and similarity information. Moreover, a suitable bounded constraint is incorporated into BRPCA to ensure that the predicted association scores locate in a meaningful interval. The extensive experiments demonstrate the superiority and robustness of the BRPCA.

中文翻译：

---

BRPCA：结合 N7-甲基鸟苷 (m7G) 位点疾病关联预测的相似性网络的有界稳健主成分分析

最近的研究表明，N7-甲基鸟苷（m ⁷ G）在各种生物过程和疾病发病机制中起着举足轻重的作用。迄今为止，转录组范围内的m ⁷ G修饰位点已通过高通量测序方法确定，一些相关信息已记录在一些生物数据库中。然而，疾病中的位点作用机制仍然未知。湿实验可以帮助高置信度地识别真正的 m ⁷ G 站点，但是在如此大量的站点中找到真正的站点非常耗时，而且成本也很高。^{因此，迫切需要计算方法来预测 m 7}之间的关联 G位点和各种疾病，从而有助于发现特定疾病的潜在活性位点。在本文中，我们提出了一种 基于相似性信息预测未知 m ^{7 G 疾病关联的有界稳健主成分分析 (BRPCA) 方法。}重要的是，BRPCA 容忍关联和相似信息中存在的噪声和冗余。此外，将合适的有界约束纳入 BRPCA 以确保预测的关联分数位于有意义的区间内。广泛的实验证明了 BRPCA 的优越性和鲁棒性。