Centrifuge: rapid and sensitive classification of metagenomic sequences

Abstract

Centrifuge is a novel microbial classification engine that enables rapid, accurate, and sensitive labeling of reads and quantification of species on desktop computers. The system uses an indexing scheme based on the Burrows-Wheeler transform (BWT) and the Ferragina-Manzini (FM) index, optimized specifically for the metagenomic classification problem. Centrifuge requires a relatively small index (4.2 GB for 4078 bacterial and 200 archaeal genomes) and classifies sequences at very high speed, allowing it to process the millions of reads from a typical high-throughput DNA sequencing run within a few minutes. Together, these advances enable timely and accurate analysis of large metagenomics data sets on conventional desktop computers. Because of its space-optimized indexing schemes, Centrifuge also makes it possible to index the entire NCBI nonredundant nucleotide sequence database (a total of 109 billion bases) with an index size of 69 GB, in contrast to k-mer-based indexing schemes, which require far more extensive space.

Deep Sequencing: Intra-terrestrial metagenomics illustrates the potential of off-grid Nanopore DNA sequencing

“…We also note the recent availability of rapid library kits including PCR-based amplification for low input samples tagged with adaptors, however the additional PCR step may co-amplify contaminants and entails further processing steps and time associated with PCR amplification and clean-up. Taxonomic classification of metagenomic reads Accordingly, about 15% of the reads (114) generated in-situ were classified by Centrifuge operating a bacterial, archaeal, viral and eukaryotic database (Kim et al 2016). As seen in Figure 3, bacterial sequences dominated the sample, with minor contributions Archaea, Eukaryota and Viruses were represented respectively by 1, 5 and 2 of the successfully classified sequences.…”

“…Bioinformatics MinION output files were processed and converted to .fastq using poRe (https://sourceforge.net/projects/rpore/, version 0.17), which was also used to generate sequencing statistics for the minION run (https://github.com/geomicrosoares). centrifuge (https://github.com/infphilo/centrifuge, version 1.0.3-beta (Kim, Song et al 2016) was run to classify passed reads using a complete bacterial, archaeal, viral and eukaryotic database (--bmax 1342177280) and pavian (https://github.com/fbreitwieser/pavian, version 0.3(Breitwieser and Salzberg 2016) ) was used to plot Sankey diagrams depicting multi-domain profiles of the microbial communities.…”

A probabilistic model to recover individual genomes from metagenomes

“…We also found that taxonomic annotation represents the most powerful information type in our simulation. For comparson, we added scores for NBC (Rosen, Reichenberger & Rosenfeld, 2011), a classifier based on nucleotide composition with in-sample training using 5-mers and 15-mers, and centrifuge (Kim et al, 2016), a similarity-based classifier both with in-sample and reference data. These programs were given the same information as the corresponding submodels and they rank close to these.…”

A novel data structure to support ultra-fast taxonomic classification of metagenomic sequences with k-mer signatures

“…These methods are often slower than BLAST alone, rendering them computationally prohibitive for large-scale analysis of many millions of short reads. However, the recent development of Centrifuge (Kim et al, 2016) has significantly improved the scalability of the alignment-based algorithm using FM-index. Besides using genomic sequences as reference, the recently published tool Kaiju (Menzel and Krogh, 2015) performs alignments towards protein sequences, achieving faster classification speed than existing tools.…”