Strbenac Dario, Zhong Ling, Raftery Mark J, Wang Penghao, Wilson Susan R, Armstrong Nicola J, Yang Jean Y H
School of Mathematics and Statistics, University of Sydney , Sydney, New South Wales 2006, Australia.
Bioanalytical Mass Spectrometry Facility, University of New South Wales , Sydney, New South Wales 2052, Australia.
J Proteome Res. 2017 Jul 7;16(7):2359-2369. doi: 10.1021/acs.jproteome.6b00882. Epub 2017 Jun 5.
Tandem mass spectrometry is one of the most popular techniques for quantitation of proteomes. There exists a large variety of options in each stage of data preprocessing that impact the bias and variance of the summarized protein-level values. Using a newly released data set satisfying a replicated Latin squares design, a diverse set of performance metrics has been developed and implemented in a web-based application, Quantitative Performance Evaluator for Proteomics (QPEP). QPEP has the flexibility to allow users to apply their own method to preprocess this data set and share the results, allowing direct and straightforward comparison of new methodologies. Application of these new metrics to three case studies highlights that (i) the summarization of peptides to proteins is robust to the choice of peptide summary used, (ii) the differences between iTRAQ labels are stronger than the differences between experimental runs, and (iii) the commercial software ProteinPilot performs equivalently well at between-sample normalization to more complicated methods developed by academics. Importantly, finding (ii) underscores the benefits of using the principles of randomization and blocking to avoid the experimental measurements being confounded by technical factors. Data are available via ProteomeXchange with identifier PXD003608.
串联质谱法是蛋白质组定量分析中最常用的技术之一。在数据预处理的每个阶段都有多种选择,这些选择会影响汇总后的蛋白质水平值的偏差和方差。利用一个新发布的满足重复拉丁方设计的数据集,开发了一套多样化的性能指标,并在一个基于网络的应用程序“蛋白质组学定量性能评估器(QPEP)”中实现。QPEP具有灵活性,允许用户应用自己的方法对该数据集进行预处理并分享结果,从而可以直接、直观地比较新方法。将这些新指标应用于三个案例研究表明:(i)肽段汇总为蛋白质的过程对所用肽段汇总方法的选择具有稳健性;(ii)iTRAQ标签之间的差异比实验批次之间的差异更强;(iii)商业软件ProteinPilot在样本间归一化方面的表现与学者们开发的更复杂方法相当。重要的是,发现(ii)强调了使用随机化和区组化原则以避免实验测量被技术因素混淆的好处。数据可通过ProteomeXchange获取,标识符为PXD003608。