A peptide-centric approach to analyse quantitative proteomics data- an application to prostate cancer biomarker discovery

Elsevier

Available online 24 November 2022, 104774

Journal of ProteomicsAuthor links open overlay panelHighlights•

New analyses of proteomics data are critical given the protein inference problem.

The p-value approach misses many relevant proteins/peptides in statistical tests.

This study showed only 24.6% consistency between protein and peptide information.

Integrated analysis of proteins and peptides offers a more complete disease view.

Peptide-centric approach reveals features that are masked during protein analysis.

Abstract

Bottom-up proteomics is a popular approach in molecular biomarker research. However, protein analysts have realized the limitations of protein-based approaches for identifying and quantifying proteins in complex samples, such as the identification of peptides sequences shared by multiple proteins and the difficulty in identifying modified peptides. Thus, there are many exciting opportunities to improve analysis methods. Here, an alternative method focused on peptide analysis is proposed as a complement to the conventional proteomics data analysis. To investigate this hypothesis, a peptide-centric approach was applied to reanalyse a urine proteome dataset of samples from prostate cancer patients and controls. The results were compared with the conventional protein-centric approach. The relevant proteins/peptides to discriminate the groups were detected based on two approaches, p-value and VIP values obtained by a PLS-DA model. A comparison of the two strategies revealed high inconsistency between protein and peptide information and greater involvement of peptides in key PCa processes. This peptide analysis unveiled discriminative features that are lost when proteins are analyzed as homogeneous entities. This type of analysis is innovative in PCa and integrated with the widely used protein-centric approach might provide a more comprehensive view of this disease and revolutionize biomarker discovery.

Significance

In this study, the application of a protein and peptide-centric approaches to reanalyse a urine proteome dataset from prostate cancer (PCa) patients and controls showed that many relevant proteins/peptides are missed by the conservative nature of p-value in statistical tests, therefore, the inclusion of variable selection methods in the analysis of the dataset reported in this work is fruitful. Comparison of protein- and peptide-based approaches revealed a high inconsistency between protein and peptide information and a greater involvement of peptides in key PCa processes. These results provide a new perspective to analyse proteomics data and detect relevant targets based on the integration of peptide and protein information. This data integration allows to unravel discriminative features that normally go unnoticed, to have a more comprehensive view of the disease pathophysiology and to open new avenues for the discovery of biomarkers.

Keywords

Prostate cancer

Mass spectrometry

Peptide-centric

Protein-centric

Urine

Biomarkers

View full text

© 2022 Elsevier B.V. All rights reserved.

留言 (0)

沒有登入
gif