---
tags: BRAILLE
title: kraken2-bracken read classification
---

[toc]

# Conda env

conda create -n kraken22 -c conda-forge -c bioconda -c defaults -c astrobiomike kraken2=2.1.1 bracken=2.6.1 bit=1.8.28

# Getting diatom genomes to put in there too:

esearch -db assembly -query '"Bacillariophyta"[Organism] AND (latest[filter] AND "representative genome"[filter] AND all[filter] NOT anomalous[filter])' | esummary | xtract -pattern DocumentSummary -def "NA" -element AssemblyAccession > assembly-accs.txt

bit-dl-ncbi-assemblies -f fasta -w assembly-accs.txt -j 5

# Making db

cat make-kraken2-and-bracken-dbs.sh

mkdir -p /data3/Data_Processing/mlee/ref-dbs/kraken2-arc-bac-vir-fungi-plant-protozoa-and-diatoms

DB="/data3/Data_Processing/mlee/ref-dbs/kraken2-arc-bac-vir-fungi-plant-protozoa-and-diatoms/"

kraken2-build --download-taxonomy --db ${DB} --threads 10

kraken2-build --download-library archaea --db ${DB} --threads 10
kraken2-build --download-library bacteria --db ${DB} --threads 10
kraken2-build --download-library viral --db ${DB} --threads 10
kraken2-build --download-library fungi --db ${DB} --threads 10
kraken2-build --download-library protozoa --db ${DB} --threads 10

for file in Bacillariophyta-NCBI-genomes-to-add-to-kraken2-db/*.fa

    kraken2-build --add-to-library ${file} --db ${DB} --threads 10

done

kraken2-build --build --db ${DB} --threads 10

bracken-build -d ${DB} -t 10 -l 250

kraken2-build --clean --db ${DB} --threads 10

# Running kraken2 and bracken

cat run-kraken2-and-bracken.sh

DB="/data3/Data_Processing/mlee/ref-dbs/kraken2-arc-bac-vir-fungi-plant-protozoa-and-diatoms/"
out_dir="kraken2-bracken-outputs-db-with-arc-bac-vir-fungi-protozoa-and-diatoms/"

mkdir -p ${out_dir}
mkdir -p ${out_dir}/kraken2-outputs/
mkdir -p ${out_dir}/bracken-outputs/

for sample in $(cat samples.txt)
do

    printf "\tOn sample: ${sample}\n\n"

    time kraken2 --db ${DB} --threads 5 --output ${out_dir}kraken2-outputs/${sample}-kraken2-out.txt --report ${out_dir}kraken2-outputs/${sample}-kraken2-report.txt --paired trimmed-data/${sample}_R1_trimmed.fq.gz trimmed-data/${sample}_R2_trimmed.fq.gz

    for rank in D P C O F G S
    do
        
        time bracken -r 250 -d ${DB} -i ${out_dir}kraken2-outputs/${sample}-kraken2-report.txt -o ${out_dir}bracken-outputs/${sample}-bracken-out-${rank}.tsv -l ${rank}
    
    done
    
done

# Formatting, summarizing, and merging kraken2 outputs
Not really going to use the bracken outputs, because I think it's being given an impossible task and the results can therefore end up wildly misleading, e.g. see [here](https://hackmd.io/@astrobiomike/tax-probing) if interested for one deep dive.

cat process-and-merge-kraken2-outputs.sh

results_dir="kraken2-bracken-outputs-db-with-arc-bac-vir-fungi-protozoa-and-diatoms/kraken2-outputs"

for sample in $(cat samples.txt)
do

    printf "\n\tWorking on sample: ${sample}\n"
    bit-kraken2-to-taxon-summaries -i ${results_dir}/${sample}-kraken2-out.txt -o ${results_dir}/${sample}-kraken2-tax.tsv

done

# combining
samples=$(sed 's/$/-kraken2-tax.tsv/' samples.txt | tr "\n" " " | sed 's/ $//')
names=$(tr "\n" "," < samples.txt | sed 's/,$//')

bit-combine-kraken2-taxon-summaries -i ${samples} -n ${names} -o All-samples-combined-kraken2-tax.tsv