Biological Function Assignment Across Taxonomic Levels in Mass-Spectrometry-Based Metaproteomics via a Modified...

Area of Science:

• Microbiology
• Bioinformatics
• Metaproteomics

Background:

• Accurate identification and quantification of microbial functions in metaproteomics is hindered by the 'shared confidently identified peptide problem'.
• Current tools often use the lowest common ancestor (LCA) algorithm, resulting in incomplete taxonomic and functional assignments.
• Existing methods struggle with precise biomass estimation and controlling false discoveries across the full microbial lineage.

Purpose of the Study:

• To develop an enhanced MiCId workflow addressing limitations in metaproteomic data analysis.
• To improve the accuracy of microorganism identification, biomass estimation, and biological function assignment.
• To provide better control over false discoveries in metaproteomic analyses.

Main Methods:

• Implementation of an expectation-maximization (EM) algorithm within the MiCId workflow.
• Integration of a comprehensive biological function database.
• Validation using synthetic datasets and reanalysis of human oral and gut microbiome data.

Main Results:

• The enhanced MiCId workflow demonstrated superior accuracy in microorganism identification and biomass estimation compared to Unipept and MetaGOmics using synthetic data.
• MiCId showed improved accuracy and better false discovery control for biological function identification versus Unipept.
• Reliable computation of function abundances across the full taxonomic lineage was achieved.
• Reanalysis of microbiome datasets yielded results consistent with original publications.

Conclusions:

• The enhanced MiCId workflow offers a significant advancement for mass-spectrometry-based metaproteomics.
• It provides more accurate and reliable identification of microorganisms and their functions, with improved false discovery control.
• This tool enhances the comprehensive understanding of microbial communities and their biological roles.