Chen Wei, Lin Hao, Feng Pengmian, Wang Jinpeng
Center for Genomics and Computational Biology, School of Sciences, Hebei United University, Tangshan, 063000, China,
Interdiscip Sci. 2014 Sep;6(3):241-9. doi: 10.1007/s12539-013-0195-4. Epub 2014 Sep 11.
Alternative splicing is a tissue and developmental stage specific process and greatly increases the biodiversity of proteins. Besides the trans- and cis-factors on the genome level, the process of RNA splicing is also regulated by epigenetic factors. In the present work, we proposed a new method to predict exon skipping events by using the histone methylation and acetylation information. The maximum relevance minimum redundancy method followed by incremental feature selection was performed to select the optimal feature set. Based on the optimized features, our method obtained an overall accuracy of 68.5% in a 10-fold cross validation test for exon skipping event prediction. It is anticipated that our method may become a useful tool for alternative splicing events prediction and the selected optimal features will provide insights into the regulatory mechanisms of epigenetic factors in alternative splicing.
可变剪接是一个组织和发育阶段特异性的过程，极大地增加了蛋白质的生物多样性。除了基因组水平上的反式和顺式因子外，RNA剪接过程还受表观遗传因子调控。在本研究中，我们提出了一种利用组蛋白甲基化和乙酰化信息预测外显子跳跃事件的新方法。采用最大相关最小冗余法并结合增量特征选择来选择最优特征集。基于优化后的特征，我们的方法在用于外显子跳跃事件预测的10折交叉验证测试中获得了68.5%的总体准确率。预计我们的方法可能成为预测可变剪接事件的有用工具，且所选择的最优特征将为表观遗传因子在可变剪接中的调控机制提供见解。
