An ONT-first bacterial isolate assembly pipeline using the best combination of fully automated tools according to the Trycycler(1) and Polypolish papers(2). Can be used with long reads alone or with optional short reads for polishing.
It takes a set of long reads and optional short reads and runs the following tools (all steps optional):
- filtlong (pre-assembly filtering) (https://github.com/rrwick/Filtlong)
- flye (assembly) (https://github.com/fenderglass/Flye)
- medaka (long-read polishing) (https://github.com/nanoporetech/medaka)
- polypolish (short-read polishing) (https://github.com/rrwick/Polypolish)
- polca (short-read polishing) (https://github.com/alekseyzimin/masurca)
#Open up a screen session
screen -R run_clinopore
#Clone the clinopore repository:
git clone https://github.com/HughCottingham/clinopore-nf.git && cd clinopore-nf
#Install the conda environments (skip this step if running on MASSIVE):
conda env create -f clinopore.yml -p /path/to/clinopore/conda/env
conda env create -f polca.yml -p /path/to/polca/conda/env
#Set run parameters:
nano nextflow.config
#Run the pipeline
conda activate /path/to/clinopore/conda/env
nextflow run clinopore.nf --clinopore_env /path/to/clinopore/conda/env --polca_env /path/to/polca/conda/envThe most important part of running the pipeline is ensuring your read data is in the correct format and has its path set as an input parameter. All reads (short and long) should be specified in the nextflow.config file (reads/*fastq.gz by default). Long reads should be in the format {isolate_id}.fastq.gz while short reads should be in the format {isolate_id}_1.fastq.gz/{isolate_id}_2.fastq.gz.
There are several other options that can be set by modifying nextflow.config. Below is a summary of these options.
| Option | Description | Default |
|---|---|---|
clinopore_env |
clinopore conda environment (MUST set) | '/path/to/clinopore/env' |
polca_env |
polca conda environment (MUST set) | '/path/to/polca/env' |
reads |
Glob pattern matching input reads | 'reads/*fastq.gz' |
medaka_model |
Medaka model (see medaka github documentation for details) | 'r941_min_sup_g507' |
filter_reads |
Optional read filtering step ('true' or 'false') | 'true' |
long_read_polish |
Optional long-read polishing step ('true' or 'false') | 'true' |
short_read_polish |
Optional short-read polishing step ('true' or 'false') | 'true' |
outdir |
Output directory | 'assemblies' |
read_qc |
Optional read QC (Nanostats and fastqc) ('true' or 'false') | 'true' |
polishing_stats |
Optional summary of bases added/subtracted at each polishing step ('true' or 'false') | 'true' |
failure_action |
Option to ignore ('ignore') or terminate ('terminate') on errors (local execution only) | 'terminate' |
threads |
Number of threads to use (local execution only) | 16 |
max_retries |
MASSIVE use only - How many times should slurm retry the job with higher CPUs, RAM and/or walltime? | 3 |
queue_size |
MASSIVE use only - Maximum number of jobs to submit at once | 50 |
slurm_account |
MASSIVE use only (MUST set if running on MASSIVE) | 'slurm_account' |
If all steps of the pipeline are run, the final assembly files will be in the base of the output directory named {isolate_id}_medaka_polypolish_polca.fasta. If you don't run every step or are interested in the intermediates, the intermediate fasta files are in the directories {output_dir}/flye, {output_dir}/medaka and {output_dir}/polypolish. The assembly graph files produced by flye are also provided in the {output_dir}/gfa directory to allow for visualisation in a tool like Bandage.
Contigs are sorted from largest to smallest, meaning the chromosome will typically be contig_1, the largest plasmid will be contig_2 and so on. Fasta headers also include contig length in bases, depth and whether it is circular or not.