Improved Prediction of Lysine Acetylation by Support Vector Machines
Improved Prediction of Lysine Acetylation by Support Vector Machines
复制标题
DOI：
10.2174/092986609788923338
复制
发表时间：
2009-08-01
影响因子：
1.6
通讯作者：
Li, Yixue
中科院分区：
文献类型：
--
作者：
Li, Songling;Li, Hong;Li, Yixue
赖氨酸残基上的可逆乙酰化是组蛋白和非组蛋白的关键翻译后修饰（PTM），控制着许多中央细胞过程。由于数据有限和缺乏明确的乙酰化共有序列，很少有研究集中在赖氨酸乙酰化位点的预测。本文综合了目前已知的赖氨酸乙酰化信息，利用支持向量机（SVM）方法沿着蛋白质序列偶联模式的编码模式，提出了一种新的赖氨酸乙酰化预测算法：LysAcet。与其他方法或现有工具相比，LysAcet是赖氨酸乙酰化的最佳预测因子，K倍（5和10）和刀切交叉验证准确率分别为75.89%，76.73%和77.16%。LysAcet的上级预测准确性主要归因于使用序列偶联模式，其描述两个氨基酸的相对位置。LysAcet有助于对赖氨酸乙酰化的有限PTM预测研究，并可作为探索蛋白质组乙酰化的补充硅内方法。在线网络服务器可在http://www.biosino.org/LysAcet/上免费获得。
Reversible acetylation on lysine residues, a crucial post-translational modification (PTM) for both histone and non-histone proteins, governs many central cellular processes. Due to limited data and lack of a clear acetylation consensus sequence, little research has focused on prediction of lysine acetylation sites. Incorporating almost all currently available lysine acetylation information, and using the support vector machine (SVM) method along with coding schema for protein sequence coupling patterns, we propose here a novel lysine acetylation prediction algorithm: LysAcet. When compared with other methods or existing tools, LysAcet is the best predictor of lysine acetylation, with K-fold (5- and 10-) and jackknife cross-validation accuracies of 75.89%, 76.73%, and 77.16%, respectively. LysAcet's superior predictive accuracy is attributed primarily to the use of sequence coupling patterns, which describe the relative position of two amino acids. LysAcet contributes to the limited PTM prediction research on lysine acetylation, and may serve as a complementary in-silicon approach for exploring acetylation on proteomes. An online web server is freely available at http://www.biosino.org/LysAcet/.