"Necklace" pipeline for creation of custom transcriptomes from RNAseq datasets, implemented in snakemake. Pipeline first published by Davidson et al. 2017 in Genome Biology. Tools used by the pipeline not included. I reccommend to use this either with the corresponding dockerfile: https://hub.docker.com/r/sarahhp/snlace/, or in another type of virtual environment (e.g. conda).
Pipeline consists of two snakemake workflows. One for preprocessing of the raw data (quality control, trimming and optional subsetting) stored in the trim_subsample.py file. The second runs the full necklace pipeline on processed data (snecklace/Snakefile), including some final summary statistics. If preprocessing is not required only run the second snakemake command in the commands.sh file.
Quick install and run on a demo dataset:
Clone this repository (not necessary if using the docker image)
git clone https://github.com/sarahhp/snlace.git
cd snlace
Create a conda environment with the necessary packages (not necessary if using the docker image)
conda env create -f snecklace/envs/nlace.yml --name nlace
conda activate nlace
Download demo data
wget https://github.com/Oshlack/necklace/wiki/asserts/sheep_small_demo_data.tar.gz
tar -C ../data -zxvf sheep_small_demo_data.tar.gz
Install extra tools (not necessary if using the docker image)
g++ -o tools/cluster tools/cluster.c
g++ -o tools/gtf2flatgtf tools/gtf2flatgtf.c
g++ -o tools/make_blocks tools/make_blocks.c
Download pblat from http://icebert.github.io/pblat/ and run make in the main dir.
Dryrun:
bash commands.sh OR snakemake -s snecklace/Snakefile -d .. --configfile sheep-demo.json --use-conda -npr
To run, remove -n option from uncommented line in commands.sh OR run snakemake -s snecklace/Snakefile -d .. --configfile sheep-demo.json --use-conda -pr
See Also:
- https://conda.io/en/latest/
- https://snakemake.readthedocs.io/en/stable/
- https://github.com/Oshlack/necklace
Usage:
- Set up an environment with the following tools and their dependencies: for both workflows:
- snakemake (at least 3.4 (I think) for snecklace pipeline)
for preprocessing workflow only:
- fastp
- seqtk
- gzip
for snecklace:
- hisat2
- samtools
- stringtie
- gffread
- trinity
- blat
- subread
- biopython
- pandas
-
Edit the config.json file with paths to data and reference genomes. To line 34 of the snecklace/scripts/get_stats.py you may have to add the absolute path of the snecklace/stats_scripts. This is because the python scripts used by get_stats are not globally installed and their location must be appended to the python path.
-
Run the pipeline by executing the commands.sh file (bash /home/commands.sh) or by running the contained snakemakemake commands directly on the CLI. -n allows a dryrun of the pipeline.