Skip to content

Quick Start

To run splicekit you need:

  1. A reference genome, downloaded and processed with pybio (installed automatically as a splicekit dependency): 
    pybio genome homo_sapiens        # or: pybio search species, for other species
  2. Aligned reads in BAM format, one file per sample. You can align FASTQ files yourself with STAR, or reuse a dataset's mapping script, e.g. datasets/GSE221868/2_map.sh, which downloads the reference genome with pybio and aligns with STAR.
  3. samples.tab — one line per sample, TAB delimited, connecting each sample_id to its treatment_id. See Sample annotation and the example samples.tab.
  4. splicekit.config — reference genome, BAM folder and the other core parameters. See Configuration and the example splicekit.config.
  5. config.yaml — per-rule Snakemake resources (cores/memory/time). Copy the template config.yaml into your project folder and adjust it to your cluster/machine.

The datasets folder has four complete examples, each with its own scripts to download and process a public RNA-seq dataset from scratch.

Running the pipeline

With samples.tab, splicekit.config and config.yaml in your project folder, run the whole pipeline with Snakemake:

cd datasets/GSE126543                               # example project folder
./1_download.sh                                     # download sample FASTQs
pybio homo_sapiens                                   # reference genome

./run_snakemake_local.sh --configfile config.yaml    # run locally
# or:
./run_snakemake_slurm.sh --configfile config.yaml    # submit jobs to SLURM

run_snakemake_slurm.sh submits each Snakemake rule as its own SLURM job (via snakemake-executor-plugin-cluster-generic), sized per-rule from config.yaml.

Once it finishes, explore the results:

splicekit web

This starts a single local web server serving both the HTML report (http://<host>:8007/report) and the JBrowse2 genome browser.

Note

If you already have BAM files and want to skip Snakemake, you can run splicekit directly with splicekit process inside a folder containing samples.tab and splicekit.config — this runs the same analysis steps sequentially on a single machine. See Command-line reference.

Next steps