Source-linked AI summary
Guidelines for the next 10 years of proteomics
Marc R Wilkins, Ron D Appel, Jennifer E Van Eyk, Maxey C M Chung, Angelika Görg, Michael Hecker, Lukas A Huber, Hanno Langen, Andrew J Link, Young-Ki Paik, Scott D Patterson, Stephen R Pennington, Thierry Rabilloud, Richard J Simpson, Walter Weiss, Michael J Dunn
TL;DR
Proteomics research faces insufficiently stringent experimental design, identification, and analytical practices. This paper develops minimum publication guidelines, establishing criteria for researchers and PROTEOMICS editors to assess proteomic manuscripts.
Problem
Proteomic studies show limited stringency in differential-expression analysis, large-scale protein identification, and complete measurement of complex peptide mixtures.
Method
The authors identified 19 issues through broad discussion and developed minimum guidelines for executing proteomics experiments and drafting manuscripts.
Results
The resulting guidelines specify requirements for replicates, biomarker performance, search parameters, false-positive estimates, and post-translational-modification evidence.
Takeaways & Limitations
The guidelines provide criteria for researchers and PROTEOMICS editors assessing the necessary and desirable characteristics of publishable proteomics manuscripts.
Takeaways & Limitations
Comprehensive proteome comparisons using these approaches require high numbers of replicates because analytical completeness varies between runs.
Abstract
from arXiv · showhide
In the last ten years, the field of proteomics has expanded at a rapid rate. A range of exciting new technology has been developed and enthusiastically applied to an enormous variety of biological questions. However, the degree of stringency required in proteomic data generation and analysis appears to have been underestimated. As a result, there are likely to be numerous published findings that are of questionable quality, requiring further confirmation and/or validation. This manuscript outlines a number of key issues in proteomic research, including those associated with experimental design, differential display and biomarker discovery, protein identification and analytical incompleteness. In an effort to set a standard that reflects current thinking on the necessary and desirable characteristics of publishable manuscripts in the field, a minimal set of guidelines for proteomics research is then described. These guidelines will serve as a set of criteria which editors of PROTEOMICS will use for assessment of future submissions to the Journal.
DIFFERENTIAL DISPLAY AND BIOMARKER DISCOVERY
Proteomic differential-display and biomarker studies require statistical consideration of analytical and biological variation, rigorous protein-identification support, and careful handling of analytical incompleteness. These concerns motivate minimum guidelines for conducting proteomics experiments and assessing manuscripts.
- Differential display and biomarker discovery: Two-fold or greater expression differences are insufficient evidence of differential expression when analytical and biological variation are ignored.Statistical tests are increasingly applied to protein-expression data, but differential expression is still frequently reported using this threshold alone.
- Protein identification: Statistical scoring systems for PMF and LC-MS/MS increasingly support high-confidence individual protein identifications.For single or small protein groups, manuscripts can also include relevant spectra or mass data.
- Analytical incompleteness: 65% overlap between replicate MudPIT identification sets demonstrates substantial analytical incompleteness in complex peptide mixtures.Thirty-five percent of proteins in the second analysis are likely to be novel relative to the first.
- Analytical incompleteness: Comprehensive proteome comparisons require great care and high numbers of replicates, while 2-D PAGE can vary in detected proteins because of sample loading and staining.Analytical completeness is also inherent to 2-D PAGE, where inconsistent loading and staining can alter the number of proteins detected in gel images.
- Moving towards new guidelines: The authors developed minimum proteomics guidelines to assist experiment execution, manuscript drafting, and editorial assessment.The guidelines were intended to establish a standard reflecting current thinking about necessary and desirable characteristics of proteomics research.
ADDENDUM
The addendum sets mandatory publication requirements for PROTEOMICS submissions, covering experimental design, statistical reporting, biomarker evaluation, protein identification, and data accessibility. Failure to follow the guidelines may lead to rejection without review.
- Experimental design and data analysis: Authors must provide experimental designs with biological and analytical replicate numbers; a single replicate is unacceptable, and clinical studies should preferably include power analysis.The guidelines also identify manuscript types suitable for PROTEOMICS in the Instructions to Authors.
- Experimental design and data analysis: Expression studies must report summary statistics, statistical results, normalization, transformation, missing-value handling, tests, degrees of freedom, and software rather than fold differences alone.Fold differences alone are explicitly deemed insufficient.
- Biomarker discovery/validation: Biomarker studies should provide sensitivity and specificity wherever possible, with ROC curves and areas under the curves considered desirable.
- Protein identification and characterization: Protein-identification reports must describe MS data generation, peak-list creation, database-search software and versions, critical parameters, searched databases, and identification-certainty measures.MS/MS reports must include peptide counts, sequences, and charge states; peptide-mass-fingerprint reports must include matching-peptide counts and total sequence coverage.
- Protein identification and characterization: Large MS/MS studies must estimate false-positive rates, while modification and isoform claims require appropriate mapping or unique-sequence evidence and supporting fragmentation data.Modification spectra and extensive identification information may be supplied as online supplementary material.
- Data and software accessibility: Academic and commercial software or databases must be freely accessible or downloadable with access options provided, and supplementary datasets should be processed rather than raw.Recommended supplementary material includes protein identifications, expression data, and MS peak lists, available online through the PROTEOMICS website.