BIOSCOPE Group, LAQV-REQUIMTE, Chemistry Department, NOVA School of Science and Technology, FCT NOVA, Universidade NOVA de Lisboa, 2829-516, Caparica, Portugal; PROTEOMASS Scientific Society, Madan Park, 2829-516, Caparica, Portugal.
Research Group of Pharmaco-Toxicological Analysis (PTA Lab), Department of Pharmacy and Biotechnology (FaBiT), Alma Mater Studiorum - University of Bologna, Via Belmeloro 6, 40126, Bologna, Italy.
Talanta. 2024 Jan 1;266(Pt 1):124953. doi: 10.1016/j.talanta.2023.124953. Epub 2023 Jul 17.
Normalization is a crucial step in proteomics data analysis as it enables data adjustment and enhances comparability between datasets by minimizing multiple sources of variability, such as sampling, sample handling, storage, treatment, and mass spectrometry measurements. In this study, we investigated different normalization methods, including Z-score normalization, median divide normalization, and quantile normalization, to evaluate their performance using a case study based on renal cell carcinoma datasets. Our results demonstrate that when comparing datasets by pairs, both the Z-score and quantile normalization methods consistently provide better results in terms of the number of proteins identified and quantified as well as in identifying statistically significant up or down-regulated proteins. However, when three or more datasets are compared at the same time the differences are found to be negligible.
归一化是蛋白质组学数据分析中的一个关键步骤，因为它可以通过最小化多个来源的变异性，例如采样、样品处理、存储、处理和质谱测量，来调整数据并增强数据集之间的可比性。在这项研究中，我们研究了不同的归一化方法，包括 Z 分数归一化、中位数划分归一化和分位数归一化，以使用基于肾细胞癌数据集的案例研究来评估它们的性能。我们的结果表明，在通过对比较数据集时，Z 分数和分位数归一化方法在鉴定和定量的蛋白质数量以及鉴定统计学上显著上调或下调的蛋白质方面始终提供更好的结果。然而，当同时比较三个或更多数据集时，差异则可以忽略不计。
