Paramer

paramer is a k-mer based sequence search tool designed to search for parasite sequences within shotgun metagenomic datasets. By leveraging Bloom filter and nthash hashing algorithm, paramer can quickly identify and classify parasitic DNA from sequencing samples. The tool is proof of concept and needs additional testing to be considered production ready.

Dependencies

• zlib
• cmake
• ntHash 2.2.0

Installation

git clone --recurse-submodules https://github.com/Ivarz/paramer
cd paramer
git -C third_party/ntHash checkout 2.2.0

Compilation

To compile:

make

To perform unit tests:

make test

Usage

Usage:
  paramer <COMMAND>

  where command is:
      mask           mask fasta file with kraken2 file and/or kmers found in another fasta file
      bloom-build    build Bloom's filter
      bloom-search   search in Bloom's filter
      extend         extend sequence in 3' and 5' directions with kmers found in Bloom's filter
      stats-bloom    gather metrics from bloom filter
      stats-fasta    gather metrics from fasta files

Tool can utilize output from kraken2 to perform reference sequence masking.

Full parasite analysis workflow consists of following steps:

1. acquire reference sequences of interest (e.g reference genomes of Ascaris lumbricoides)

wget https://ftp.ebi.ac.uk/pub/databases/wormbase/parasite/releases/WBPS19/species/ascaris_lumbricoides/PRJEB4950/ascaris_lumbricoides.PRJEB4950.WBPS19.genomic.fa.gz

2. perform reference classification with kraken2. For gut microbiome sequencing datasets I recommend using host specific kraken2 catalogue (if available). Assuming kraken2 is available in $PATH:

kraken2 \
        --threads 24 \
        --output k2.out.txt \
        --report k2.rep.txt \
        --db /path/to/kraken2/database \
        ascaris_lumbricoides.PRJEB4950.WBPS19.genomic.fa.gz

3. perform reference masking from the obtained kraken2 results:

paramer mask \
    -f ascaris_lumbricoides.PRJEB4950.WBPS19.genomic.fa.gz \
    -k k2.out.txt \
    -o ascaris_lumbricoides.PRJEB4950.WBPS19.genomic.masked.fa.gz

4. build bloom filter from the reference

paramer bloom-build \
    -k 31 \
    -w 31 \
    -s 10G \
    -r ascaris_lumbricoides.PRJEB4950.WBPS19.genomic.masked.fa.gz \
    -o ascaris_lumbricoides.PRJEB4950.WBPS19.genomic.masked.blm

5. search your sequencing data against Bloom filter. Assuming paired-end sequencing sample names sample_1.fq.gz and sample_2.fq.gz:

paramer bloom-search \
        -b ascaris_lumbricoides.PRJEB4950.WBPS19.genomic.masked.blm \
        -i sample_1.fq.gz \
        -I sample_2.fq.gz \
        -o sample_output_1.fq.gz \
        -O sample_output_2.fq.gz

Will output files containing sequencing reads that have matches in the Bloom filter (in example: sample_output_1.fq.gz and sample_output_2.fq.gz)