Centrifuge

Centrifuge classifies DNA sequences from microbial samples rapidly and with low memory usage to provide taxonomic labels and species abundance estimates for metagenomic datasets.

Key Features:

• Indexing scheme: Combines the Burrows-Wheeler transform (BWT) with the Ferragina-Manzini (FM) index to create compressed searchable indices.
• Space-optimized indices: Implements space-optimized data structures that reduce index size for large genome collections.
• Low memory footprint: Can create an index for 4,078 bacterial and 200 archaeal genomes using approximately 4.2 GB of storage.
• Large-database indexing: Can index the NCBI nonredundant nucleotide sequence database (109 billion bases) with an index size of about 69 GB.
• High-throughput classification: Processes millions of reads from high-throughput DNA sequencing runs in minutes for taxonomic labeling.
• Species quantification: Produces estimates of species presence and relative abundance from classified reads.
• Efficiency versus k-mer methods: Achieves substantially smaller index sizes compared with traditional k-mer-based indexing methods.

Scientific Applications:

• Metagenomic taxonomic classification: Label sequencing reads from complex microbial communities including bacteria and archaea.
• Species abundance profiling: Quantify species presence and relative abundance in metagenomic samples.
• Large-scale reference searches: Classify reads against large reference collections such as the NCBI nonredundant nucleotide sequence database.
• High-throughput sequencing analysis: Analyze millions of reads generated by high-throughput DNA sequencing runs for taxonomic and abundance information.

Methodology:

Combines the Burrows-Wheeler transform (BWT) with the Ferragina-Manzini (FM) index and space-optimized data structures to build compact indices for genome sets and the NCBI nonredundant nucleotide sequence database, then performs read classification and species-level quantification.