Snakemake workflow: rna-seq-star-tetranscripts

A Snakemake workflow for transposable element RNA-Seq using TEtranscripts.

Usage

Install Snakemake

Snakemake is best installed via Conda/Mamba:

conda create -n snakemake -c conda-forge -c bioconda snakemake>=8.25.5
conda activate snakemake

Deploy workflow

Clone the repository:

git clone https://github.com/niekwit/rna-seq-star-tetranscripts.git
cd rna-seq-star-tetranscripts

Place your raw reads (.fastq.gz) in the reads/ directory. For single-end data, files should be named {sample}.fastq.gz. For paired-end data, files should follow the pattern {sample}_R1_001.fastq.gz / {sample}_R2_001.fastq.gz.

Configure workflow

config/config.yaml

Parameter	Description
genome	Genome build (mm39, hg38, hg19, mm38, T2T-CHM13v2.0)
ensembl_genome_build	Ensembl release number (e.g. 113)
strand	Library strandedness: forward, reverse, or unstranded (see TEtranscripts docs)
fdr_cutoff	FDR threshold for volcano plots (default: 0.05)
fc_cutoff	log₂ fold change threshold for volcano plots (default: 0.5)
spike_in.apply	Set to True to enable spike-in normalisation
spike_in.name	Pattern matching spike-in feature names (e.g. ERCC-)
spike_in.gtf	Path to spike-in GTF file
spike_in.fasta	Path to spike-in FASTA file
mapping.outFilterMultimapNmax	Max number of multi-mapping loci allowed (default: 100)
mapping.winAnchorMultimapNmax	Max multi-mapping loci for anchor windows (default: 100)
mapping.extra_params	Any additional STAR arguments
split_bam	Split BAMs by chromosome before TE counting — recommended for large datasets on a cluster
deeptools.binsize	Bin size (bp) for bigWig generation
deeptools.normalisation	Normalisation method: RPKM, FPKM, or TPM

config/samples.csv

A comma-separated file describing the samples:

sample,genotype,treatment,reference,batch
WT_1,WT,None,yes,1
KO_1,KO,None,no,1

Column	Description
sample	Unique sample name matching the reads filename (without .fastq.gz)
genotype	Genotype of the sample
treatment	Treatment condition
reference	Set to yes for control samples used as the reference in differential expression
batch	Sequencing batch. If more than one batch is present, a batch effect is modelled in DESeq2

Run workflow

Locally (Conda environments resolved automatically):

snakemake --cores <N> --use-conda

On an HPC with Slurm:

snakemake --cores <N> --use-conda --slurm --default-resources slurm_account=<account> slurm_partition=<partition>

With Apptainer (using the pre-built Docker image):

snakemake --cores <N> --use-apptainer

To perform a dry run first:

snakemake -n

Output

Results are written to the results/ directory:

Path	Description
results/qc/	FastQC and MultiQC reports
results/trimmed/	Adapter-trimmed reads
results/mapped/	STAR-aligned BAM files and per-sample mapping stats
results/te_count/	TEcount count tables (genes + TEs)
results/bigwig/	RPKM-normalised bigWig tracks
results/deseq2/	DESeq2 results CSVs for each comparison (genes and TEs separately)
results/plots/	Volcano plots, PCA, sample distance heatmap, mapping rates, and TE class plots

Citation

If you use this workflow in a paper, don't forget to give credits to the authors by citing the URL of this (original) repository and its DOI (see above).