Nanopore nextflow pipeline

Clinical pipeline for real time analysis of ONT sequencing data.

This pipeline includes species identification (using Centrifuge) and resistance gene identification (using RGI-CARD).

This pipeline assumes the use of a multiplexing sequencing kit as it will gather fasta files based on the barcodes provided :

The following inputs are expected :

• fastq : the folder in which sequencing data were gathered (absolute path, it should end by "/barcodeXX", e.g. XX = 01)
• sampleID : a unique name for your sample
• summary : a basecalling summary file from guppy (absolute path)

an optionnal input can be provided to try to use reads for which demultiplexing failed with --keep_unclassified.

• unclassified : a single ONT fastq file containing reads for which demultipexing failed

These information have to be gather in a csv formated file, with the following header (in any order): sampleId,fastq,summary(,unclassified)

A complete run looks like the following : nextflow run nanopore_main.nf --samples <your_input_file.csv> --outdir <output_folder_name> --summary --tax

Currently, the whole pipeline needs to be run to create a multiQC report (found in the folder multiqc). But you can also search for tool specific outputs in folders : centrifuge, resistance, coverage or checkM.

Parameters to modify

A few parameters have to be modified to fit your databases, the simplest way is to update the nextflow.config file :

• human_genome (provide an assembly of the human genome)
• pepper_db (location of pepper's database, only needed for the evaluation pipeline)
• homopolish_db (location of the homopolish's mash_sketches)
• cendb (location of the centrifuge database)
• card (location of the CARD database)

Polishing tools evaluation

To reproduce the polishing evaluation results, you can use the nanopore_evaluation.nf pipeline.

The inputs expected are a little different from the above pipeline :

• fastq : a single ONT fastq file. (absolute path, ´cat´ all passed reads is a way to create such a file)
• sampleID : a unique name for your sample
• summary : a basecalling summary file from guppy (absolute path)
• reference : a reference genome (absolute path)

These information have to be gather in a csv formated file, with the following header (in any order): sampleId,fastq,summary,reference

A complete run looks like the following : nextflow run nanopore_evaluation.nf --samples <your_input_file.csv> --outdir <output_folder_name> --summary --map --coverage --tax --reference --res --annotation

Currently, the whole pipeline needs to be run to create a multiQC report (found in the folder multiqc). But you can also search for tool specific outputs in folders : centrifuge, resistance, coverage or checkM.