Considerations and Software for Successful Immune Cell Deconvolution Using Proteomics Data.

Area of Science:

• Proteomics
• Computational Biology
• Immunology

Background:

• Bulk transcriptomics is common for cell-type deconvolution.
• Proteomics data has been underutilized for cell deconvolution.
• Estimating cell composition offers biological insights.

Purpose of the Study:

• To evaluate computational methods for immune cell deconvolution using bulk proteomics data.
• To assess the impact of preprocessing and software on deconvolution accuracy.
• To introduce a tool for proteomics-based cell deconvolution.

Main Methods:

• Utilized defined immune cell populations and simulated mixtures.
• Assessed various preprocessing techniques and software tools.
• Developed and validated an R package named proteoDeconv.

Main Results:

• Demonstrated the feasibility of cell-type deconvolution using proteomics data.
• Achieved Pearson correlations above 0.9 for estimated proportions in simulated mixtures.
• Identified optimal parameters for missing value imputation and reference matrix generation.

Conclusions:

• Proteomics data is a viable alternative for cell-type deconvolution.
• The proteoDeconv R package simplifies proteomics data preprocessing for deconvolution.
• This approach enhances the analysis of cell composition in biological samples.