mashID

Identify organisms from genome assemblies (fasta) and long or short raw reads (fastq).

Description

MashID was developed to quickly identify bacterial species from WGS data. Users can provide their own mash sketches database to identify other organisms.

MashID works with Illumina (paired-end or single-end), Ion Torrent, Nanopore and PacBio data. Actually any fastq (sequencing data) or fasta (genome assemblies), gzipped or not, can be used as input.

MashID comes with a pre-compiled database to identify Mycobacteria species. Note that this database doesn't perform well with Mycobacterium tuberculosis variant bovis. Please use the new database in Figshare (link at bottom) for this identifying Mycobacteria.

Installation

# Create and activate virtual envrionment
conda create -n mashID -y -c bioconda python=3.10 mash=2.3 pandas=1.4.4 psutil=5.9.2 bbmap=38.99
conda activate mashID

# Clone repo and test mashID
git clone https://github.com/duceppemo/mashID
cd mashID
python mashID.py -h

Usage

usage: python mashID.py [-h] -i /input/folder/ -o /output/folder/ [-d /path/to/mash_databse.msh] [--identity 0.9] [--p-value 0.05] [-n 10]
                        [-s {identity,multiplicity}] [-t 64] [-p 2] [-m 459] [-v]

Species identification from NGS data using Mash.

options:
  -h, --help            show this help message and exit
  -i /input/folder/, --input /input/folder/
                        Input directory with fastq/fasta files or a single fastq/fasta file, gzipped or not.
  -o /output/folder/, --output /output/folder/
                        Output directory
  -d /path/to/mash_databse.msh, --database /path/to/mash_databse.msh
                        Mash sketch database. Will run a Mycobacteria mash database by default.
  --identity 0.9        Minimum identity to report. [0-1]. Default is 0.9
  --p-value 0.05        Maximum p-value to report
  -n 10, --n-hits 10    Number of top-hits to report (sorted by % identity). Default is 10.
  -s {identity,multiplicity}, --sort-by {identity,multiplicity}
                        How to sort the result tables. Will impact the "Top hit" table. Default is "similarity". Optional.
  -t 64, --threads 64   Number of threads. Default is maximum available(64). Optional.
  -p 2, --parallel 2    Number of samples to process in parallel. Default is 2. Optional.
  -m 459, --memory 459  Memory in GB. Default is 85% of total memory (459)
  -v, --version         show program's version number and exit

Outputs

• sample1_mashID.tsv: individual sample mash screen output table.
• summary_mashID.tsv: if more than one sample, this file will hold the top identification result for each sample.

Building custom database

You can use the make_mashID_db.py script to build a suitable database for mashID.

python make_mashID_db.py \
    -i /media/data/sample_folder \
    -o /media/database/mycoID' \

Run python make_mashID_db.py -h for detailed help.

• Pre-compiled databases for Refseq bacteria and proGenomes v3, Listeria spp. and Mycobacteria spp. can be found here