蛋白质定量是探索疾病发生发展状况和寻找新药靶标的重要手段。在shotgun蛋白组学中,目前常用定量方法包括综合同位素标记后的质谱峰强度方法和无标记定量方法。根据数据类型无标记定量方法可以分为两类:基于鉴定蛋白的质谱数的方法和基于质谱峰强度的方法。本研究主要用EM算法改进基于鉴定蛋白质谱数的定量方法,并用免疫印迹实验获得的酵母全蛋白的丰度来验证EM算法改进后定量的有效性。结果表明,改进后的质谱数和蛋白丰度的相关性比改进前有一定的提高。同时,利用这些数据对主要的几种基于鉴定蛋白的质谱数的模型进行了比较,发现PAI模型最好,SpS模型次之,emPAI模型最不适合于蛋白质定量。
Protein quantification is one of the most important methods to explore the diseased state or developmental stages, which would be valuable for developing new drug targets for various diseases. The most common techniques in shotgun proteomics include the integration of mass spectrometric peak intensities with stable isotope labeling and the label-free protein mass spectrometric (MS) quantification. The latter is mainly classified into two categories : spectral count of identified proteins and MS peak area intensity of identified peptides. This paper mainly discussed the method using expectation maximitation (EM) -like algorithm to improve the spectral count of identified proteins. Then the method was evaluated with yeast proteome data sets. The results show that EM-like algorithm is effective, and the Pearson correlation between spectral counts of proteins and protein abundances are improved after using EM-like algorithm. We also compared the three models for protein quantification using the yeast data sets, and found that protein abundance index (PAI) model is the best in these methods, spectrum sampling (SpS) the second and exponentially moditied protein abundance index do last.