Species classifier choice is a key consideration when analysing low-complexity food microbiome data

Area of Science:

• Microbiology
• Bioinformatics
• Genomics

Background:

• Shotgun metagenomics is valuable for analyzing food microbial communities.
• No consensus exists on optimal species classification tools, platforms, or sequencing depth.
• Benchmarking is needed to evaluate high-throughput sequencing platforms for food microbiota analysis.

Purpose of the Study:

• To benchmark Illumina MiSeq, NextSeq 500, and Ion Proton platforms for shotgun metagenomics of food microbiota.
• To assess the impact of sequencing depth and bioinformatic tools on analysis results.
• To compare species classification accuracy, false-positive rates, and metagenome assembly.

Main Methods:

• Sequencing of six kefir DNA samples and one mock community DNA sample.
• Analysis using various bioinformatic tools for species classification and functional profiling.
• Testing the effects of sequencing depth by random subsampling of reads.

Main Results:

• Sequencing platforms showed consistent compositional analysis results across different depths.
• Species classification tools significantly impacted results, with MetaPhlAn2 showing the lowest false-positive rate.
• Metagenome assembly completeness was lower on MiSeq compared to NextSeq and Proton, improving with sequencing depth.

Conclusions:

• Sequencing platforms demonstrated similar performance in shotgun metagenomics of food microbiota.
• Bioinformatic methodology choice had a greater impact on results than the sequencing platform.
• Further optimization of bioinformatic tools is crucial for accurate food microbiota analysis.