mdBG

mdBG constructs a minimizer-space de Bruijn graph representation to enable efficient genome assembly and large-scale genomic analyses from long-read sequencing data.

Key Features:

• Efficiency: Achieves orders-of-magnitude improvements in computational efficiency, assembling a human genome in under 10 minutes using 8 CPU cores and 10 GB of RAM and assembling 60 Gbp of metagenomic reads in 4 minutes with 1 GB of RAM.
• Scalability: Scales to population-scale datasets, demonstrated by constructing a minimizer-space de Bruijn graph representation for 661,405 bacterial genomes comprising 16 million nodes and 45 million edges.
• Metagenomics and Pangenomics: Enables processing and representation of very large and diverse genomic datasets to support metagenomic and pangenomic analyses.
• Anti-Microbial Resistance (AMR) Gene Detection: Has been applied to search for AMR genes within the graph representation, completing the search in 12 minutes.

Scientific Applications:

• Genome Assembly: Assembly of genomes from long-read sequencing data using minimizer-space de Bruijn graphs to reduce computational resources.
• Metagenomics: Analysis and assembly of large metagenomic datasets containing diverse microbial communities.
• Pangenomics: Comparative assembly and analysis across multiple genomes for pangenomic studies.
• AMR Gene Detection: Searching for antimicrobial resistance genes within large graph-based genomic representations.

Methodology:

Constructs minimizer-space de Bruijn graphs by extracting minimizers as representative subsequences and building the de Bruijn graph in minimizer-space to reduce graph complexity and size.