MGAP can now use SPAdes instead of Velvet if needed.
MGAP has been updated to version 2.0 with Nextflow integration for job management. Installation and running the pipeline is now slightly different than previous versions.
MGAP Installation (post version 2.0)
MGAP has a large number of dependencies that need to be installed prior to running the pipeline and there are a number of tools shipped with the pipeline itself. To install MGAP, first download the github repository into a local directory called mgap.
git clone https://github.com/dsarov/MGAP.git ./mgap
Add execute permissions for the tools shipped with MGAP
chmod -R +x ./mgap/*
Install the environment with Conda, although I can almost never get conda to solve environments on my system and prefer to use Mamba.
conda env create --name mgap -f ./mgap/env.yaml
If you have issues with conda, you can try using mamba instead
conda install -c conda-forge mamba
mamba env create --name mgap -f ./mgap/env.yaml
To run the pipeline, copy (or link) paired-end Illumina reads into a working directory and (optionally) suppy a closely related reference for scaffolding. Activate the MGAP environment in Conda
conda activate mgap
And then the main section of the pipeline can be run by:
nextflow run /path_to_mgap/main.nf
Nextflow Integration
As of version 2.0, MGAP utilises nextflow for job management. This means that pipeline management is now slightly different to previous versions. Cluster options and pipeline options are now set in the nextflow.config file. Multiple job management systems are now supported (e.g. SGE, slurm, PBS etc) and can be specified by the --executor flag in the pipeline. The pipeline can now be resumed by the use of the -resume flag when initiating the pipeline.
What is MGAP?
MGAP (Microbial Genome Assembler Pipeline) is an automated pipeline that runs an optimised Velvet assembly followed by gap filling and polishing to produce high-quality, reference-assisted assembly of microbial genomes using paired-end Illumina data.
MGAP workflow
MGAP performs reference assisted assemblies on paired-end Illumina FASTQ reads, either in phred +33 or phred +64 quality score format. MGAP cannot be used on single-end data, or for NGS data generated on platforms other than Illumina.
To achieve high-quality assemblies, MGAP incorporates the following programs into its workflow:
-
Trimmomatic (Bolger et al., 2014)
-
SPAdes (Bankevich et al., 2012)..
-
Velvet (Zerbino & Birney, 2008)
-
VelvetOptimiser (https://github.com/tseemann/VelvetOptimiser)
-
GapFiller (Boetzer & Pirovano, 2012)
-
ABACAS (Assefa et al., 2009)
-
IMAGE (Tsai et al., 2010)
-
SSPACE (Boetzer et al., 2011)
-
Pilon (Walker et al., 2014)
-
MIRA (convert_project) (Chevreux et al., 1999)
How do I install MGAP (prior to version 2.0)?
- Download or clone the MGAP git repository.
git clone https://github.com/dsarov/MGAP---Microbial-Genome-Assembler-Pipeline.git
- Add executable permissions to the installation
cd MGAP---Microbial-Genome-Assembler-Pipeline
chmod -R +x ./*
-
Edit the MGAP.config to point to the installation location of MGAP. You should just need to change the MGAP_LOCATION variable.
-
Test out the pipeline. MGAP should complain if it can't find any of its dependencies.
How do I run MGAP?
USAGE: MGAP.sh -r [reference, without .fasta extension] -s [specify single strain]
If you would prefer MGAP to perform assemblies without using a reference to assist or if no reference is available, set the -r flag to "none".
MGAP expects reads to be paired-end Illumina data in the following format: StrainOne_1_sequence.fastq.gz, StrainOne_2_sequence.fastq.gz (first pair) and StrainTwo_1_sequence.fastq.gz, StrainTwo_2_sequence.fastq.gz (second pair).
What does MGAP do?
Prior to assembly, FASTQ reads are conservatively trimmed and filtered using Trimmomatic v0.35 to remove low-quality bases and Illumina adapter contamination. The Trimmomatic parameters in MGAP are: LEADING=3, TRAILING=3, SLIDINGWINDOW=4:15, MINLEN=36, and ILLUMINACLIP (for TruSeq2 paired-end adapters). These parameters can be altered as desired.
Next, a draft scaffold assembly is created using Velvet 1.2.10, with parameters optimised using VelvetOptimiser v2.2.4 at a default kmer range of 53 to 75; these kmers can also be altered if required.
GapFiller v2.1.1 is then used to attempt to fill in the scaffolds created by Velvet. Following creation of the draft Velvet assemblies, the best assembly is then improved upon using ABACAS v1.3.1 and IMAGE v2.4.1. If the user provides a reference genome, ABACAS scaffolds the contigs against this reference. If no user-specified reference is provided, this ABACAS step is skipped. Scaffolded contigs are then attempted to be joined using IMAGE, which will break contigs that have been incorrectly joined. A second attempt is then made to scaffold contigs using SSPACE v3.0. GapFiller is then run again to attempt to fill in the scaffolds created by SSPACE.
Finally, Pilon v1.22 corrects any insertion-deletion (indel) and SNP errors in the final assembly. Contigs <1,000bp are excluded from the final .fasta assembly output using the miraconvert tool in MIRA v4.
Who created MGAP?
MGAP was written by Derek Sarovich (@DerekSarovich) and Erin Price (@Dr_ErinPrice).
What to do if I run into issues with MGAP?
Please send bug reports to [email protected]