HoCoRT

HoCoRT removes host-derived sequences from shotgun metagenomic sequencing data to reduce bias, computational load, and privacy risk and enable accurate downstream microbial community analyses.

Key Features:

• Genome indexing: Provides a one-step option for genome indexing of host reference sequences.
• Methodological flexibility: Incorporates mapping, classification, and alignment methods and allows selection of underlying classifiers and parameter adjustment.
• Performance recommendations: Includes evaluated speed and accuracy trade-offs with recommended methods for short reads such as BioBloom, Bowtie2 (end-to-end), and HISAT2, and notes that Kraken2 offers highest speed with reduced accuracy.
• Long-read decontamination: Supports long-read datasets by combining Kraken2 with Minimap2, reporting high accuracy and detection of 59% of human reads in long-read tests.
• Comparative performance: Using Bowtie2 in end-to-end mode, demonstrates faster processing and slightly improved accuracy compared to DeconSeq.

Scientific Applications:

• Metagenomic decontamination: Removal of host sequences to prevent skewing of taxonomic and functional analyses in shotgun metagenomes.
• Human microbiome studies: Decontamination of human gut and oral microbiome sequencing data to reduce host-derived bias and privacy concerns.
• Environmental and complex samples: Applicable to samples containing host contamination to improve accuracy of community profiling.
• Short- and long-read sequencing: Applicable across short-read and long-read sequencing technologies for host sequence removal.

Methodology:

Performs genome indexing and identifies/removes host reads using mapping, classification, and alignment programs explicitly including BioBloom, Bowtie2 (end-to-end), HISAT2, Kraken2, and Minimap2, with benchmarking of speed and accuracy across methods.