Pangebin is a pipeline aimed at performing the task of "plasmid binning" on a bacterial sample
exploiting a pangenome construction on a pair of assembly graph using nf-core/pangenome,
running a modified version of @chauvec's Plasbin-Flow.
Requirements
- nextflow + java 17
- python 3.10
- gurobi license
- conda/mamba
Prior to run the pipeline
Clone the repoistory, cd into it and checkout the ismb2025 branch:
git clone ...
cd pangebin
git checkout ismb2025
Prepare your samples: provide for each sample a folder named after its biosample_id,
inside put the gfa(s) built for your short read sample using Unicycler and Skesa, named
respectively as unicycler.gfa.gz and skesa.gfa.gz.
Note that the gfa files are compressed using bgzip (from htslib).
The dataset used for the experiment used for ISMB2025 submission can be found in the dataset.tar.gz file.
De-compress it, the command will create a database folder.
tar xavf dataset.tar.gz
Running the pipeline
Run the pipeline on a sample of choice in the dataset (replace xxxxxx with an actual folder name found in dataset).
export SAMPLE_ID="xxxxxxxx"
Chose either conda or mamba as running backend, to run the main pangebin pipeline:
nextflow run . --input dataset/$SAMPLE_ID -profile conda
or
nextflow run . --input dataset/$SAMPLE_ID -profile mamba
To run also original plasbin-flow on skesa and unicycler assemblies separately run:
nextflow run . --input dataset/$SAMPLE_ID -profile conda --assemblers
or
nextflow run . --input dataset/$SAMPLE_ID -profile mamba --assemblers
You can find the results in the same input folder:
- binning output
- plasbin-flow
$SAMPLE_ID.{ske,uni}.1.pbf.bins.tsv - pangebin
$SAMPLE_ID.pan.1.pbf.bins.tsv
- plasbin-flow
- binning results
- plasbin-flow
$SAMPLE_ID.{ske,uni}.1.pbf.pred.{ske,uni}.bin.txt - pangebin:
$SAMPLE_ID.pan.1.pbf.pred.{uni,ske}.bin.txt
- plasbin-flow
- labeling results
- plasbin-flow
$SAMPLE_ID.{ske,uni}.1.pbf.pred.{ske,uni}.lab.txt - pangebin
$SAMPLE_ID.pan.1.pbf.pred.{uni,ske}.lab.txt
- plasbin-flow