STRONG

STRONG resolves strains within metagenomic datasets by leveraging assembly graphs, coassembly, and binning into metagenome-assembled genomes (MAGs) to identify de novo strains across multiple metagenome samples.

Key Features:

• Assembly graph-based resolution: Stores the coassembly graph prior to variant simplification to enable strain-resolving analysis on graph structures.
• Coassembly and binning into MAGs: Integrates coassembly and binning to produce metagenome-assembled genomes for downstream strain analysis.
• SCG-focused subgraph extraction: Extracts subgraphs and obtains unitig per-sample coverages for individual single-copy core genes (SCGs) within each MAG.
• BayesPaths Bayesian inference: Applies the BayesPaths Bayesian algorithm to infer the number of strains, their haplotypes (sequences on SCGs), and their abundances.
• Per-sample unitig coverage: Uses unitig per-sample coverage information to inform strain and haplotype estimation across samples.

Scientific Applications:

• De novo strain identification: Identifies strains across multiple metagenome samples without prior reference genomes.
• Haplotype reconstruction on SCGs: Reconstructs strain haplotypes specifically on single-copy core genes within MAGs.
• Strain abundance estimation: Estimates abundances of inferred strains across samples using unitig coverages and Bayesian inference.
• Benchmarking with synthetic communities: Validated strain-identification accuracy using synthetic community datasets.
• Time-series and long-read comparison: Applied to anaerobic digester time series and produced haplotypes that closely matched long Nanopore reads.

Methodology:

Performs coassembly and binning into MAGs, stores the coassembly graph before variant simplification, extracts subgraphs and unitig per-sample coverages for individual SCGs within each MAG, and applies the BayesPaths Bayesian algorithm to infer the number of strains, SCG haplotypes, and their abundances.