Send to

Choose Destination
J Proteomics. 2019 Jul 10;207:103441. doi: 10.1016/j.jprot.2019.103441. [Epub ahead of print]

Five simple yet essential steps to correctly estimate the rate of false differentially abundant proteins in mass spectrometry analyses.

Author information

Univ. Grenoble Alpes, CEA, INSERM, BIG-BGE, 38000 Grenoble, France.
Bioinformatics and Biostatistics Hub, C3BI, Institut Pasteur, USR 3756 IP CNRS, 75015 Paris, France; Proteomics Platform, Mass Spectrometry for Biology Unit, Institut Pasteur, USR 2000 IP CNRS, 75015 Paris, France.
Univ. Grenoble Alpes, CEA, INSERM, BIG-BGE, 38000 Grenoble, France; CNRS, BIG-BGE, F-38000 Grenoble, France. Electronic address:


Results from mass spectrometry based quantitative proteomics analysis correspond to a subset of proteins which are considered differentially abundant relative to a control. Their selection is delicate and often requires some statistical expertise in addition to a refined knowledge of the experimental data. To facilitate the selection process, we have considered differential analysis as a five-step process, and here we present the practical aspects of the different steps. Prostar software is used throughout this article for illustration, but the general methodology is applicable with many other tools. By applying the approach detailed here, researchers who may be less familiar with statistical considerations can be more confident in the results they present.


Data processing; Differential analysis; False discovery rate; Quantitative proteomics; Statistical software

Supplemental Content

Full text links

Icon for Elsevier Science
Loading ...
Support Center