Strainberry: automated strain separation in low-complexity metagenomes using long reads

Area of Science:

• Microbiology
• Bioinformatics
• Genomics

Background:

• High-throughput short-read metagenomics enables species-level analysis but struggles with strain resolution.
• Microbial strains within a species can exhibit significant functional differences.
• Long-read sequencing technologies offer potential for accurate genome assembly and strain differentiation.

Purpose of the Study:

• To introduce Strainberry, a novel metagenome assembly pipeline.
• To achieve strain separation in metagenomic data using exclusively long-read sequencing.
• To evaluate Strainberry's performance on both mock and real microbial communities.

Main Methods:

• Development of Strainberry, a long-read based metagenome assembly pipeline.
• Benchmarking on mock communities to assess assembly accuracy and completeness.
• Application on real-world metagenomic datasets to evaluate performance on complex samples.

Main Results:

• Strainberry produced near-complete, highly accurate (99.9% base accuracy) strain-resolved assemblies on mock communities.
• Application on real datasets yielded 20-118% more genomic material for individual strain genomes compared to conventional methods.
• Demonstrated successful separation of strain genomes in a complex microbiome, refining diversity assessment.

Conclusions:

• Strainberry effectively performs strain separation in low-complexity metagenomes using long-read data.
• The pipeline significantly improves the recovery and resolution of individual strain genomes from complex microbial communities.
• This work provides a foundation for advancing strain-resolved metagenome assembly in more complex environments.