Biological Function Assignment across Taxonomic Levels in Mass-Spectrometry-Based Metaproteomics via a Modified Expectation Maximization Algorithm

A major challenge in mass-spectrometry-based metaproteomics
is
accurately identifying and quantifying biological functions across
the full taxonomic lineage of microorganisms. This issue stems from
what we refer to as the “shared confidently identified peptide
problem″. To address this issue, most metaproteomics tools
rely on the lowest common ancestor (LCA) algorithm to assign biological
functions, which often leads to incomplete biological function assignments
across the full taxonomic lineage of identified microorganisms. To
overcome this limitation, we implemented an expectation-maximization
(EM) algorithm, along with a biological function database, within
the MiCId workflow. Using synthetic datasets, our study demonstrates
that the enhanced MiCId workflow achieves better control over false
discoveries and improved accuracy in microorganism identification
and biomass estimation compared to Unipept and MetaGOmics. Additionally,
the updated MiCId offers improved accuracy and better control of false
discoveries in biological function identification compared to Unipept,
along with reliable computation of function abundances across the
full taxonomic lineage of identified microorganisms. Reanalyzing human
oral and gut microbiome datasets using the enhanced MiCId workflow,
we show that the results are consistent with those reported in the
original publications, which were analyzed using the Galaxy-P platform
with MEGAN5 and the MetaPro-IQ approach with Unipept, respectively.