The selection of a data processing method for use in mass spectrometry-based label-free proteome quantification contributes significantly to its accuracy and precision. In this study, we comprehensively evaluated 7 commonly-used label-free quantification methods (MaxQuant-Spectrum count, MaxQuant-iBAQ, MaxQuant-LFQ, MaxQuant-LFAQ, Proteome Discoverer, MetaMorpheus, TPP-StPeter) with a focus on missing values, precision, accuracy, selectivity, and reproducibility of low abundance protein quantification in both single shot and fractionation. Our results showed that among the tested strategies, MaxQuant in MaxLFQ mode outperformed other strategies in terms of accuracy and precision in both whole proteome and low abundance proteome quantification, whereas the Proteome Discoverer (PD) strategy using SEQUEST as a search engine performed better in terms of quantifiable low abundance proteome coverage. We subsequently applied the PD and MaxLFQ strategies in a blood proteomic dataset and found that many FDA-approved tumor prognostic biomarkers could be identified as well as quantified using the PD strategy, indicating the potential advantage of PD in label-free quantification studies. These results provide a reference for method choice in label-free quantification data analysis. SIGNIFICANCE: Mass spectrometry-based label-free quantification methods play an important role in label-free proteome data analysis. In this study, we evaluated 7 commonly-used label-free quantification methods with respect to the following aspects: missing values, precision, accuracy, selectivity, and reproducibility for low abundance protein quantification. The results showed that, among the strategies evaluated, the PD strategy with SEQUEST as a search engine performed better in terms of low abundance protein coverage. This study provides a reference for method choice in label-free quantification data analysis.
Keywords: Accuracy; Label-free quantification; Precision; Proteomics.
Copyright © 2020 Elsevier B.V. All rights reserved.