TaxIt: An Iterative Computational Pipeline for Untargeted Strain-Level Identification Using MS/MS Spectra from...

Area of Science:

• Microbiology
• Bioinformatics
• Proteomics

Background:

• Accurate strain-level classification of unknown organisms via mass spectrometry is difficult.
• Existing reference databases lack taxonomic depth, limiting identification to species level.
• Larger databases increase runtime and decrease statistical power due to proteome similarity.

Purpose of the Study:

• To present TaxIt, an iterative workflow for MS/MS-based strain-level classification of samples with unknown taxonomic origins.
• To address the challenges of large search spaces and ambiguous taxonomic assignments in microbial identification.
• To enhance the confidence and accuracy of taxonomic classification at the strain level.

Main Methods:

• TaxIt employs an iterative workflow starting with reference sequence data for species identification.
• Automated acquisition of relevant strain sequences facilitates low-level classification.
• An abundance weighting strategy resolves ambiguous assignments caused by proteome similarities.

Main Results:

• TaxIt successfully identified microbial strains in bacterial and viral samples.
• The iterative workflow demonstrated superior performance compared to noniterative approaches.
• The method achieved correct strain identification in all tested examples, with one tie.

Conclusions:

• TaxIt offers a robust solution for untargeted, strain-level taxonomic classification.
• The workflow enhances the potential for deeper and more accurate microbial identification.
• TaxIt leverages public sequence resources and is available as open-source software.