Hostile

Hostile removes human host sequences from microbial sequencing data to prevent host-derived contamination and preserve microbial reads for accurate downstream analyses such as variant calling and de novo assembly.

Key Features:

• High Accuracy in Sequence Removal: Removes at least 99.6% of real human reads from datasets.
• Retention of Microbial Sequences: Retains ≥99.989% of simulated bacterial reads, increasing to ≥99.997% with a masked reference genome while decreasing human read removal by ≤0.001%.
• Performance and Efficiency: Removes 21%–23% more human short reads than comparator tools while misclassifying 21–43 times fewer bacterial reads and typically requires less processing time.
• Versatility with Input Data: Accepts paired and unpaired fastq[.gz] input files and supports sequencing reads ranging from short to long.

Scientific Applications:

• Clinical microbial genomics: Removes host contamination from clinical metagenomic samples to enable accurate microbial analyses.
• Variant Calling: Improves the accuracy of microbial variant calling by eliminating confounding human DNA.
• De Novo Assembly: Enhances de novo assembly of microbial genomes by reducing host-derived noise.

Methodology:

Computational algorithms and pipeline steps are not specified in the provided description.