MetaAll: integrative bioinformatics workflow for analysing clinical metagenomic data
Martin Bosilj1, Alen Suljič1, Samo Zakotnik1
1Institute of Microbiology and Immunology, Faculty of Medicine, University of Ljubljana, Zaloška cesta 4, 1000 Ljubljana, Slovenia.
Briefings in Bioinformatics
|November 16, 2024
Summary
A new bioinformatics workflow, MetaAll, enhances pathogen detection in clinical metagenomics. This method improves sensitivity and specificity, aiding in identifying infectious agents in complex cases.
Area of Science:
- Microbiology
- Bioinformatics
- Genomics
Background:
- Metagenomics has advanced significantly, yet a universal bioinformatics pipeline for clinical applications remains elusive.
- Current methods struggle with sensitivity and specificity in pathogen detection from complex metagenomic data.
Purpose of the Study:
- To develop and validate a user-friendly bioinformatics workflow (MetaAll) for enhanced clinical metagenomics.
- To improve pathogen detection sensitivity and specificity using a multi-step approach.
Main Methods:
- Combined three bioinformatics tools into a three-step workflow named MetaAll.
- Analyzed short paired-end (PE) and long reads from metagenomic datasets.
- Validated the workflow using four complex clinical cases and the CAMI Clinical pathogen detection challenge dataset.
Main Results:
- MetaAll successfully identified putative pathogens in most clinical cases, including co-infections (e.g., Haemophilus influenzae/Human rhinovirus, SARS-Cov-2/Influenza A).
- Identified Klebsiella pneumoniae in a case where conventional diagnostics failed.
- Demonstrated improved pathogen detection compared to traditional methods in challenging samples.
Conclusions:
- The MetaAll workflow offers a robust solution for clinical metagenomic analysis.
- Combining read classification, contig validation, and targeted reference mapping is crucial for reliable infectious agent detection.
- MetaAll shows significant potential for improving diagnostics in clinical microbiology.
