This pipeline uses snakemake, minimap2, salmon, edgeR, DEXSeq and stageR to automate simple differential gene expression and differential transcript usage workflows on long read data.

If you have paired samples (e.g for example treated and untreated samples from the same individuals) use the paired_dge_dtu branch.

The input files and parameters are specified in config.yml:

• transcriptome - the input transcriptome.
• annotation - the input annotation in GFF format.
• control_samples - a dictionary with control sample names and paths to the fastq files.
• treated_samples - a dictionary with treated sample names and paths to the fastq files.

• alignments/*.bam - unsorted transcriptome alignments (input to salmon).
• alignments_sorted/*.bam - sorted and indexed transcriptome alignments.
• counts - counts generated by salmon.
• merged/all_counts.tsv - the transcript count table including all samples.
• merged/all_counts_filtered.tsv - the transcript count table including all samples after filtering.
• merged//all_gene_counts.tsv - the gene count table including all samples.
• de_analysis/coldata.tsv - the condition table used to build model matrix.
• de_analysis/de_params.tsv - analysis parameters generated from config.yml.
• de_analysis/results_dge.tsv and de_analysis/results_dge.pdf- results of edgeR differential gene expression analysis.
• de_analysis/results_dtu_gene.tsv, de_analysis/results_dtu_transcript.tsv and de_analysis/results_dtu.pdf - results of differential transcript usage by DEXSeq.
• de_analysis/results_dtu_stageR.tsv - results of the stageR analysis of the DEXSeq output.
• de_analysis/dtu_plots.pdf - DTU results plot based on the stageR results and filtered counts.

• miniconda - install it according to the instructions.
• snakemake install using conda.
• pandas - install using conda.
• The rest of the dependencies are automatically installed using the conda feature of snakemake.

• README.md
• Snakefile - master snakefile
• config.yml - YAML configuration file
• snakelib/ - snakefiles collection included by the master snakefile
• lib/ - python files included by analysis scripts and snakefiles
• scripts/ - analysis scripts
• data/ - input data needed by pipeline - use with caution to avoid bloated repo
• results/ - pipeline results to be commited - use with caution to avoid bloated repo

Clone the repository:

git clone https://github.com/nanoporetech/pipeline-transcriptome-de.git

Edit config.yml to set the input datasets and parameters then issue:

snakemake --use-conda -j <num_cores> all

(c) 2018 Oxford Nanopore Technologies Ltd.

This Source Code Form is subject to the terms of the Mozilla Public License, v. 2.0. If a copy of the MPL was not distributed with this file, You can obtain one at http://mozilla.org/MPL/2.0/.

This pipeline is largely based on the approach described in the following paper:

• Love MI, Soneson C and Patro R. Swimming downstream: statistical analysis of differential transcript usage following Salmon quantification. F1000Research 2018, 7:952 (doi: 10.12688/f1000research.15398.3)

See the post announcing the tool at the Oxford Nanopore Technologies community here.