RNAja
About RNAja
RNAja is a pipeline written in snakemake, allowing to analyse RNAseq data, perform differential expression analysis using several mappers and counters. RNAja is the only pipeline able to compare method effects over DE genes.
1. Install dependencies and clone RNAja
Check dependencies for RNAja : Python < 3.12.0 and Snakemake >= 7.32.4
Here, you can install RNAja from the gitlab source code such as :
git clone git@forge.ird.fr:phim/rnaja-pipeline.git
cd rnaja-pipeline
python3 -m pip install .
# in edition mode (for developpers and debuging)
#python3 -m pip install -e .
Don't forget to export RNAja into the $PATH environment path by
export PATH=/home/$USER/.local/bin/:$PATH`
1.1 Installing in local mode
RNAja install_local --help
RNAja install_local
1.2 Installing in cluster mode
Install RNAja in cluster mode using singularity container from rnaja_utilities https://itrop.ird.fr/RNAja_utilities/
RNAja install_cluster --help
RNAja install_cluster --scheduler slurm --env singularity
Please run command line 'RNAja edit_cluster_config' before the first run of RNAja see https://forge.ird.fr/phim/rnaja-pipeline/README.md and configure by default threads and memory resources.
Feel free to modify resources (RAM, CPU number and partition) using the rules names found above.
rule hisat2_index
rule hisat2_map
rule star_index
rule star_map_count
rule star_to_diffex
rule samtools_stats
rule multiqc
rule stringtie_discovery
rule stringtie_gtf_list_discovery
rule merge_stringtie_gtf_discovery
rule stringtie
rule stringtie_gtf_list
rule merge_stringtie_gtf
rule list_for_prepDE
rule prepDE_stringtie_table
rule htseq_count
rule htseq_to_diffex
rule diff_exp_analysis
In the cluster_config.yaml
file adapt ressources as in this example adapted:
_default__:
cpus-per-task: 4
mem-per-cpu: 3G
partition : normal
output: '{log.output}_cluster'
error: '{log.error}_cluster'
job-name: '{rule}.{wildcards}'
rule stringtie_discovery
nodelist: node4
cpus-per-task: 8
mem-per-cpu: 6G
partition: highmem
2. Running a datatest
Running a test to be sure of RNAja installation. RNAja test_install
download an available dataset from a distant server (RNAja_utilities i-Trop). Pipeline tests are launched in a repertory (option -d DIR_NAME) created by RNAja.
RNAja test_install --help
RNAja test_install -d TEST
2.1 Running your data in LOCAL mode
launching suggested command line done by RNAja, in LOCAL mode:
RNAja run_local --help
RNAja run_local -t 8 -c TEST/data_test_config.yaml --singularity-args "--bind $HOME"
In local mode, it's possible to allocate threads to some rules using --set-threads
snakemake argument such as
RNAja run_local -t 8 -c TEST/data_test_config.yaml --set-threads hisat2_index=4 hisat2_map_paired=2
2.2 Running your data in CLUSTER mode
Now you can launch suggested command line done by RNAja in cluster
mode BUT before you need to configurate cluster parameters running command line RNAja edit_cluster_config
before the first run and modify threads, ram, node and computer resources.
RNAja does a copy of cluster_config.yaml
file into your home "/home/$USER/.config/rnaja_pipeline/cluster_config.yaml"
RNAja run_cluster --help
RNAja edit_cluster_config
If singularity was selected in installation of RNAja, it could be needed to give argument --singularity-args \"--bind $HOME\"
to RNAja, by using :
RNAja run_cluster --help
RNAja run_cluster -c TEST/data_test_config.yaml --singularity-args "--bind $HOME"
# here a example adapted to @IFB HPC
RNAja run_cluster -c TEST/data_test_config.yaml --singularity-args "--bind /shared:/shared"
#you can also use snakemake parameters as
RNAja run_cluster -c TEST/data_test_config.yaml --singularity-args "--bind $HOME" --rerun-incomplete --nolock
Important Note : In IRD "i-Trop" cluster, run RNAja using ONLY a node, data has to be in "/scratch" of chosen node. Use nodelist : nodeX
parameter inside of cluster_config.yaml
̀ file.
3. Running your data
config.yaml
file
3.1.create a pipeline by configuring the Before to run RNAja, create a config.yaml
by using before to adapt it.
RNAja create_config
Three sections are needed for RNAja into the config.yaml
file: section DATA, MODE and PARAMS.
DATA section
Adapt config.yaml
file with path to fastq files, reference and annotation file in the DATA section
as well as the output directory.
DATA:
fastq_dir: "/path/to/FASTQ"
reference: "/path/to/reference/ref.fna"
annotation: "/path/to/reference/ref.gtf"
output_dir: "RNAJA_OUTPUT"
MODE section
Five pipelines can be running in parallel by using RNAja !
We have included for instance two mappers (STAR and HISAT2) and 3 counters (STRINGTIE, HTSEQCOUNT,STAR). You can activate or deactivate pipelines you would run in the MODE section such as ...
MODE:
HISAT2_STRINGTIE: true
HISAT2_HTSEQCOUNT: true
STARmap_STARcount: true
STARmap_HTSEQCOUNT: true
STARmap_STRINGTIE: true
PARAMS section
In the PARAMS section, tools parameters can be modified and adapted.
PARAMS:
HISAT2:
indexation:
prefix: "REF"
STAR:
indexation:
params: "--sjdbOverhang 100 --genomeSAindexNbases 10"
mapping:
params: "--readFilesCommand zcat" # --outFilterMismatchNoverLmax 0.03
STRINGTIE:
discovery_mode : false
HTSEQCOUNT:
params: "-r pos -s reverse -m union -t gene "
You can modify HISAT2
indexation, STAR
indexation and mapping options, change STRINGTIE
mode (discovery or not) as well as HTSEQCOUNT
params. Feel free to check documentation of these tools before to run RNAja!
output
Here an example of output_dir
if you have activated all five pipelines proposed by RNAja.
RNAJA_OUTPUT/
├── COUNT
│ ├── HTSEQCOUNT
│ │ ├── HISAT2
│ │ └── STAR
│ ├── STAR
│ │ ├── Batch-rep1Aligned.out.bam
│ │ ├── Batch-rep1Aligned.toTranscriptome.out.bam
│ │ ├── Batch-rep1Log.final.out
│ │ ├── Batch-rep1Log.out
│ │ ├── Batch-rep1Log.progress.out
│ │ ├── Batch-rep1ReadsPerGene.out.tab
│ │ ├── Batch-rep1SJ.out.tab
...
│ │ ├── CENPK-rep1Aligned.out.bam
│ │ ├── CENPK-rep1Aligned.toTranscriptome.out.bam
│ │ ├── CENPK-rep1Log.final.out
│ │ ├── CENPK-rep1Log.out
│ │ ├── CENPK-rep1Log.progress.out
│ │ ├── CENPK-rep1ReadsPerGene.out.tab
│ │ ├── CENPK-rep1SJ.out.tab
...
│ └── STRINGTIE
│ ├── HISAT2_Batch-rep1.gtf
│ ├── HISAT2_Batch-rep1.tsv
...
│ ├── HISAT2_CENPK-rep3.gtf
│ ├── HISAT2_CENPK-rep3.tsv
│ ├── STAR_Batch-rep1.gtf
│ ├── STAR_Batch-rep1.tsv
...
│ ├── STAR_CENPK-rep3.gtf
│ ├── STAR_CENPK-rep3.tsv
├── LOGS
├── MAPPING
│ ├── HISAT2
│ │ ├── Batch-rep1.bam
│ │ ├── Batch-rep1.bam.csi
│ │ ├── Batch-rep1_HISAT_summary.txt
...
│ │ ├── CENPK-rep3.bam
│ │ ├── CENPK-rep3.bam.csi
│ │ └── CENPK-rep3_HISAT_summary.txt
│ └── STAR
│ ├── Batch-rep1.bam
│ ├── Batch-rep1.bam.csi
...
│ ├── CENPK-rep3.bam
│ └── CENPK-rep3.bam.csi
├── REF
│ ├── HISAT2
│ │ ├── GCF_000146045.2_R64_genomic.fasta -> /scratch/rnaja_test/TEST/DATA_TEST/REF/GCF_000146045.2_R64_genomic.fna
│ └── STAR
│ ├── chrLength.txt
│ ├── chrNameLength.txt
│ ├── chrName.txt
│ ├── chrStart.txt
│ ├── exonGeTrInfo.tab
│ ├── exonInfo.tab
│ ├── GCF_000146045.2_R64_genomic.fasta -> /scratch/rnaja_test/TEST/DATA_TEST/REF/GCF_000146045.2_R64_genomic.fna
│ ├── geneInfo.tab
│ ├── Genome
│ ├── genomeParameters.txt
│ ├── Log.out
│ ├── SA
│ ├── SAindex
│ ├── sjdbInfo.txt
│ ├── sjdbList.fromGTF.out.tab
│ ├── sjdbList.out.tab
│ └── transcriptInfo.tab
└── slurm_logs
Authors
- Aurore Comte (IRD), Christine Tranchant (IRD), Julie Orjuela (IRD)
Some parts of RNAja code and documentation were inspired or came from the pipelines below.
-
Culebront (Julie Orjuela et al.) https://github.com/SouthGreenPlatform/culebrONT
-
sRNAmake (Sebastien Cunnac et al.) https://github.com/Aucomte/sRNAmake
-
BulkRNA (Camille Cohen) https://github.com/CamilleCohen/ProjetTuteur-_BulkRNA
RNAja uses really nice python package SnakEcdysis https://snakecdysis.readthedocs.io/en/latest/package.html to perform installation and execution in local and cluster mode. SnakEcdysis is developed by Sébastien Ravel (CIRAD).
Acknowledgements
Thanks to Ndomassi Tando (i-Trop IRD) by administration support.
The authors acknowledge the IRD i-Trop HPC https://bioinfo.ird.fr/ (South Green Platform http://www.southgreen.fr) at IRD Montpellier for providing HPC resources that have contributed to this work.
Thanks to Alexis Dereeper for his help and the development of diffexDB https://bioinfo-web.mpl.ird.fr/cgi-bin2/microarray/public/diffexdb.cgi.
License
Licenced under MIT https://opensource.org/license/mit/. Intellectual property belongs to IRD and authors.