AltSS

Alternative Splicing Snakemake pipeline, to process data from RNA-Seq to alternative splicing prediction. Only working with rMATS [1] and SplAdder [2] tools so far.

[1]https://github.com/Xinglab/rmats-turbo
[2]https://spladder.readthedocs.io/en/latest/index.html

Requirements

Tested in WSL2 and linux server

• Conda 24.11.3
• GTF and STAR index files for respective genomic release (Genecode or Ensembl).

1.1. Prepare rMATS environment

1. Download rmats-turbo in a folder with the same name.

   git clone https://github.com/Xinglab/rmats-turbo.git

2. Create conda environment for Snakemake and rMATS.

   conda env create -n snakemake -f envs/snakemake.yml
   conda env create -n rmats -f envs/rmats.yml

3. Activate rmats environment, and compile rmats:

   conda activate rmats
   cd rmats-turbo/
   ./build_rmats

1.2. Prepare SplAdder environment

1. Create conda environment for Snakemake and SplAdder.

   conda env create -n snakemake -f envs/snakemake.yml
   conda env create -n spladder -f envs/spladder.yml

2. Prepare RNA-seq files

To correctly run the snakemake file, prepare each of the filenames inside the RNA-seq data in the follow way:

For paired reads

GlobalSampleName_GroupName*_SampleNumber_1.fastq.gz
GlobalSampleName_GroupName*_SampleNumber_2.fastq.gz

For single reads

GlobalSampleName_GroupName*_SampleNumber.fastq.gz

*Group name corresponds to the condition and control samples.

Example

HEK293_wt_1_1.fastq.gz      HEK293_wt_1_2.fastq.gz
HEK293_mut_1_1.fastq.gz     HEK293_mut_1_2.fastq.gz
...

3. Configure Pipeline

Edit config.yaml with the following variables:

Required Parameters

• raw_path: Directory path to RNA-seq raw data to be used.
• gtf: Path to genomic GTF reference file.
• fasta: Path to genome FASTA file (required for generating a new STAR index).
• outdir: Output directory for results.
• read_length: Read length of your sequencing data. Default: 50.

Optional Parameters

• read_type: Either "paired" or "single". Default: "paired".
• control_name: Name of the control group for identification of the group in the name file
• nthread: Number of threads to use. Default: 8 cores.

STAR Index Generation

The pipeline can automatically generate a STAR index. This requires significant RAM (typically 30-40GB for human genome). Ensure your system has sufficient memory or use lower number of cores (nthread).

4. Run snakemake

Activate snakemake environments and run snakemake

conda activate snakemake
snakemake -c --use-conda -s rMATS-pipeline.smk --configfile config.yaml

Or for SplAdder:

conda activate snakemake
snakemake -c --use-conda -s SplAdder-pipeline.smk --configfile config.yaml

NOTE: Very first run of the SnakeMake file, it will initialize new conda environments. This may take a several minutes.

Troubleshooting

ERROR: libgsl.so.25: cannot open shared object file: No such file or directory

If rMATS finishes with the aforementioned error, it means that the program is not locating the correct path for that function. Modify this by using the follow bash command, while in the snakemake environment.

export LD_LIBRARY_PATH=path/to/rmats_pipeline/.snakemake/conda/environement_created_for_rmats/lib