Strategies to enable large-scale proteomics for reproducible research

Area of Science:

• Proteomics
• Analytical Chemistry
• Computational Biology

Background:

• Reproducible research is crucial for scientific advancement.
• Large-scale proteomics requires robust and reproducible mass spectrometry (MS) methods.
• Variability in MS data can arise from instrument differences and time-dependent factors.

Purpose of the Study:

• To assess the reproducibility of data-independent acquisition mass spectrometry (DIA-MS) over time and across instruments.
• To develop computational methods for improving quantitative accuracy and reducing variability in large-scale proteomics.
• To establish a computational pipeline for enhancing the analysis of DIA-MS data for clinical applications.

Main Methods:

• Performed 1560 DIA-MS runs using eight different sample types on six mass spectrometers over four months.
• Utilized negative controls and replicates to identify and remove unwanted variation.
• Developed computational modules for reducing missing values and mitigating instrument-specific variations.

Main Results:

• The developed computational methods significantly improved quantitative accuracy and reduced variability compared to existing approaches.
• The ProNorM pipeline effectively mitigated variations among instruments over time.
• Accurate prediction of tissue proportions was achieved, demonstrating the utility of the methods.

Conclusions:

• The study provides a computational framework (ProNorM) to enhance the reproducibility and quantitative accuracy of large-scale DIA-MS.
• These advancements offer a viable pathway for the deployment of proteomics in clinical settings.
• Improved data analysis methods are essential for realizing the full potential of high-throughput proteomics.