预测和鉴定蛋白质翻译后修饰的生物信息方法

蛋白质翻译后修饰对蛋白质成熟、结构和功能多样性有决定性的作用。但蛋白质翻译后修饰的多样性、普遍性、动态性,使传统的生物化学方法在全局水平上理解翻译后修饰非常有限,对它们的研究、特别是大规模的研究长期发展缓慢。现在,在实验研究基础上,借助多方面的生物信息学方法,可以快速高通量的预测和鉴定蛋白质翻译后修饰。一方面,可以从序列角度出发,基于酶识别底物的特异性,用位点权重矩阵、支持向量机等算法,从底物蛋白质序列提取修饰相关的保守序列,并用于预测翻译后修饰位点。这种方法相对成熟,能够取得较理想的预测准确性,但不能反映不同时间不同细胞的翻译后修饰状态。另一方面,可从质谱数据分析出发,有望捕获细胞内翻译后修饰的动态特性。质谱分析的高灵敏度、高准确度和高通量的能力已使建立在质谱基础上的蛋白质组学成为研究翻译后修饰的重要工具,生物信息学方法和质谱蛋白质组学的结合则更可以加速研究翻译后修饰的进程。本文从序列和质谱分析两个角度总结评价了各种翻译后修饰相关生物信息学方法的研究近况,重点讨论利用质谱数据鉴定翻译后修饰的新思路。

Bioinformatics Methods for Predicting and Identifying Post-Translational Modifications

Post-Translational Modifications (PTMs) are crucial to protein maturation, protein structure and protein function. PTMs exist in vivo ubiquitously with dynamic diversity, making it very difficult to globally study them with traditional biochemistry methods. Large-scale PTM study has been slow in progress. Recently, the development of new proteomics technologies, with the assis- tance of various bioinformatics methods, has made it possible to predict and identify PTMs in a high-throughput way. In one way, based on the binding specificity of enzymes, multiple methods can be applied to extract PTM-related conserved motifs from substrate protein sequences, and then to predict PTM sites from new sequences. Such methods like position-specific scoring matrix and support vector ma- chine learning etc. are relatively mature and often acquire a good accuracy, but they can not capture dynamic PTMs of various cells in different states. In another way, the quick progress of mass spectrometry technology has rendered it an important tool to study PTMs be- cause of its high sensitivity, accuracy and high-throughputness. It becomes possible to capture dynamic PTMs by coupling proteomics technology and bioinformatics. This review summarizes recent applications of bioinformatics methods on studying PTMs from these two aspects, with an emphasis on the identification of PTMs using mass spectrometry.