transXpress-nextflow

transXpress: a Nextflow pipeline for rapid de novo transcriptome assembly and annotation

Also see our sister project: transXpress-snakemake

Intro

Dependencies

Requires

• NextFlow 20.04.0+ (install via conda)
• fastqc (install via conda)
• trimmomatic (install via conda)
• Trinity (install via conda)
• SPAdes (install via conda)
• TransDecoder (install via conda)
• BioPython (install via conda)
• samtools (install via conda)
• bowtie2 (install via conda)
• infernal (install via conda)
• HMMER (install via conda)
• kallisto (install via conda)
• NCBI BLAST+ (install via conda)
• R (install via conda)
• seqkit (install via conda)
• basic Linux utitilies: wget, split, awk, cut, gzip

Optional

• deeploc / SignalP 4.1 / SignalP 5.0
• tmhmm v. 2.0

Installation

1. Install Miniconda3
2. Setup conda environment

conda create --name transxpress-nf
conda activate transxpress-nf

3. Install conda dependencies:

 conda config --add channels bioconda
 conda config --add channels conda-forge
 conda config --add channels r
 conda config --set channel_priority false
 conda install -y nextflow fastqc trimmomatic "trinity>=2.13.2" "spades>=3.15.4" "transdecoder>=5.5.0" biopython samtools bowtie2 infernal hmmer kallisto blast r r-tidyverse seqkit bioconductor-edger parallel graphviz

(Note, below dependencies are optional, transXpress will run to completion without them, but will produce empty files for their output)

4. Install deeploc (performance being evaluated by transXpress developers in comparison to SingalP 4.1/5.0)

   • Download deeploc from http://www.cbs.dtu.dk/cgi-bin/nph-sw_request?deeploc
   • Install dependencies: pip install -r requirements.txt
   • Install deeploc: python setup.py install or locally: python setup.py install --user

5. Install SignalP 4.1g (performance being evaluated by transXpress developers in comparison to SingalP 5.0/deeploc)

   • Download SignalP 4.1g from http://www.cbs.dtu.dk/cgi-bin/sw_request?signalp+4.1

6. Install SignalP 5.0 (performance being evaluated by transXpress developers in comparison to SingalP 4.1/deeploc)

   • Download SignalP 5.0 from http://www.cbs.dtu.dk/cgi-bin/nph-sw_request?signalp

7. Install tmhmm

   • Download tmhmm from http://www.cbs.dtu.dk/services/TMHMM/

Usage

Make your assembly directory and change it to the current directory

mkdir your_assembly_directory
cd your_assembly_directory

Setup the mandatory 'samples.tsv' file in the assembly directory describing where to find your raw read FASTQ files. Reads will be pooled from all samples for a single transcriptome assembly, but expression quantification will happen on a per-sample basis. See the tests directory for an example of a samples file: samples.tsv

Setup the mandatory 'prefix.txt' file in the directory describing which genus species the data comes from, or whichever metadata you prefer to add. See the tests directory for an example of a species file: prefix.txt

Symbolically link the transxpress-nextflow code into your assembly directory

ln -s /your/transxpress-nextflow-cloned-directory/* ./

Make sure your conda transxpress environment has been sourced, and then execute the run.sh script with your assembler and profile of choice. You can choose your execution/cluster platform by setting the --executor parameter, e.g. local or pbs Note: For the cluster, depending on how strict your cluster is, you may need to tweak cluster.config quite a bit.

nextflow run main.nf -w work-$ASSEMBLER -profile $THEPROFILE --assembler $ASSEMBLER --samples 'samples.tsv' --prefix_add_metadata_file 'prefix.txt' -resume

NextFlow only likes 1 assembly per directory, so if you'd like to run two assemblies simultaneously, you have to use different assembly directories.

Assemblers

Currently 'trinity' or 'rnaspades'

Profiles

The 2nd parameter for the ./run.sh wrapper script allows you to specify the profile that is used. The profiles (stored in the nextflow.config file) are currently used to configure the execution mode (cluster vs local), and if the assembly is strand specific or not.

./run.sh trinity strandSpecific

Available profiles are as follows:

strandSpecific
notStrandSpecific
test_notStrandSpecific
test_strandSpecific

Running tests

cd ./tests/
cd ./test_nonSS-trinity ##non strand specific assembly using trinity. Other directories have other assemblers / parameters.
./run_test.sh

Flow graph