Benchmarking MicrobIEM - a user-friendly tool for decontamination of microbiome sequencing data

Area of Science:

• Microbiology
• Bioinformatics
• Genetics

Background:

• Microbiome analysis using 16S rRNA gene sequencing is common in research.
• Contaminants in sequencing data can distort results, particularly in low-biomass environments like human skin.
• Bioinformatic removal of contaminants is crucial for accurate microbiome analysis.

Purpose of the Study:

• Introduce MicrobIEM, a novel tool for bioinformatically removing contaminants from 16S rRNA gene sequencing data using negative controls.
• Evaluate MicrobIEM's performance against existing decontamination methods.
• Provide a user-friendly tool for microbiome quality control.

Main Methods:

• Benchmarking MicrobIEM against five established decontamination approaches.
• Utilizing four 16S rRNA amplicon sequencing datasets: three mock communities with varying contamination levels and compositions, and one skin microbiome dataset.
• Employing evaluation measures such as Youden's index for unbiased assessment.

Main Results:

• Decontamination performance varied based on algorithm parameters and dataset type.
• Control-based algorithms, including MicrobIEM's ratio filter, were effective in staggered mock communities and low-biomass samples.
• MicrobIEM and Decontam's prevalence filter successfully reduced contaminants in the skin microbiome dataset while preserving relevant genera.

Conclusions:

• MicrobIEM's ratio filter is a highly effective decontamination method, comparable to established tools.
• MicrobIEM offers a user-friendly graphical interface with interactive plots for parameter selection.
• MicrobIEM is the first quality control tool designed for microbiome researchers without coding expertise.