mjsull/ngpipe

Pipeline for typing Neisseria spp

ngpipe

ngpipe is a pipeline for typing Neisseria gonorrhoeae. It takes a folder of reads or contigs and runs sequence typing for MLST, NG-STAR and NG-MAST typing. It also determines species using the rplF gene and checks for the presence of rplF and ppng. It will compile this information into a table and create a minimum spanning tree using MLST.

Installation

Clone this repository

git clone https://github.com/mjsull/ngpipe.git

Requirements

Essential

Python

networkx

matplotlib

Snakemake - https://github.com/snakemake/snakemake

MLST - https://github.com/tseemann/mlst

Minimap

Samtools

Spades

Optional

Abricate

FastQC

Usage

To run ngpipe on a list of contigs

1. Create text file /path/to/samples.txt
2. Enter a list of samples to process, one on each line.
3. create directory /path/to/contigs
4. For each sample copy a FASTA of contigs to /path/to/contigs/<sample>.fasta
5. Run

snakemake --snakefile ngpipe/Snakefile --config workdir=/path/to/output contig_dir=/path/to/contigs samples=/path/to/samples.txt --cores 4

To run ngpipe on a folder of Illumina reads

1. Create text file /path/to/samples.txt
2. Enter a list of samples to process, one on each line.
3. create directory /path/to/reads
4. For each sample copy read pairs to to /path/to/reads/<sample>_R1.fastq.gz and /path/to/reads/<sample>_R2.fastq.gz
5. Run

snakemake --snakefile ngpipe/Snakefile --config workdir=/path/to/output read_dir=/path/to/reads samples=/path/to/samples.txt --cores 4

Implementation

ngpipe first checks for database updates to the NG-STAR, NG-MAST MLST and rplF schemes from pubMLST and downloads them to the MLST directory. It then builds a new index for the software "mlst". If this is the first time running ngpipe they will be downloaded from scratch.

If running on read sequences, reads are assembled using Spades. This pipeline is intended to be used with amplicons and thus parameters are used to primarily assemble amplicons well which may result in suboptimal assemblies if run on isolate read data.

Assembled contigs are then typed using Torsten Seeman's software: mlst. Plasmid-mediated resistance in gonorrhoea is detected via amplification of a region encmpassing the plasmid and the blaTEM gene. This region is detected by mapping reads or contigs to a reference.. The output is then compiled. Some additional processing is done to the output to account for bugs in the mlst software (for example calling multiple alleles when a stop loss occurs).

Finally a minimum spanning tree will be drawn using networkx.