The purpose of this pipeline is to remove host-associated reads from a pathogen sequencing project. This may be done to meet privacy and ethics requirements when analyzing samples that have been collected from human hosts, or as a quality control process to ensure that host-derived reads do not interfere with downstream processes such as genome assembly.
This pipeline performs the following tasks:
- Estimates the relative abundances of reads originating from a host organism and a pathogen of interest using kraken2 and bracken
- Performs 'de-hosting' by aligning reads against a reference genome using bwa, and extracting unmapped reads using samtools
- Verifies that the remaining reads do not contain host-derived sequenced by again estimating relative abundances using the same method as in step 1.
- Prepares a brief report summarizing the efficacy of the de-hosting process.
| Parameter | Default Value | Description |
|---|---|---|
kraken2_db |
Path to kraken2 database | |
bracken_db |
Path to bracken database | |
host_reference |
Path to host reference genome | |
taxonomy_level |
'S' |
Taxonomic level at which to group reads ('S' = Species) |
read_length |
150 |
Input sequence read length. Must match bracken database. |
host_name |
'Homo sapiens' |
Name of host. Must match name in kraken2 database. |
pathogen_name |
'Severe acute respiratory syndrome-related coronavirus' |
Name of pathogen of interest. Must match name in kraken2 database. |
Build a kraken2 database.
kraken2-build --db <path_to_kraken2_db> --standard
Build the bracken database.
bracken-build -d <path_to_kraken2_db> -l <read_length>
Index the host reference genome.
bwa index <path_to_host_reference>