scKenver is a comprehensive reproducible benchmarking framework for long-read single-cell and spatial transcriptomics.
It provides four independent and modular Nextflow workflows:

1. CompareToShortReads
   Benchmark gene-level quantification and cell metrics using Illumina as reference.

2. UMI_transcript_assignment
   Evaluate UMI–transcript mapping quality and error-correction.

3. Isoform_discovery
   Assess structural accuracy of isoform predictions from different tools.

4. Comparison_with_ground_truth
   Compare results against a known ground truth (simulated or validated datasets).

Each workflow can run individually or through a global orchestrator (main.nf in repository root).

Repository structure:

scKenver/
├── main.nf                         # Global orchestrator
├── nextflow.config                 # Global configuration
├── 1-CompareToShortReads/
├── 2-UMI_transcript_assignment/
├── 3-Isoform_discovery/
├── 4-Comparison_with_ground_truth/
├── Matrix_samplesheet.csv          # input sheet for WF1
├── GTF_samplesheet.csv             # input sheet for WF3
└── bin/

A complete, zipped test dataset is available on Zenodo:

Example download:

wget -O scKenver_testdata.zip "https://zenodo.org/records/17632476/files/test_data.zip?download=1"
unzip scKenver_testdata.zip -d data_test

• Nextflow ≥ 21.10.3
• Docker (recommended) or Singularity/Apptainer
• Git

Setup:

git clone https://github.com/alihamraoui/scKenver.git
cd scKenver
nextflow -version

The root main.nf orchestrates all workflows end-to-end.

nextflow run main.nf -profile docker   --outdir results_global   -with-report -with-trace -with-timeline

Outputs:

• results_global/metrics/ — consolidated benchmark results
• results_global/figures/ — QC plots + benchmark visualizations
• execution report — Nextflow logs, trace, timeline

Purpose: Compare gene/cell metrics between long-read and short-read datasets.

Steps:

• Matrix building & normalization
• Concordance of cell/gene metrics
• QC + visualization
• Spatial metrics if spatial data

Command:

cd 1-CompareToShortReads
nextflow run main.nf --samplesheet ../Matrix_samplesheet.csv   --outdir ../results_wf1

Outputs:

• correlation matrices, UMI/gene/cell recovery metrics
• QC reports

Purpose: Evaluate UMI ↔ transcript assignment quality.

Command:

cd 2-UMI_transcript_assignment
nextflow run main.nf  --data_raw ../data_test/umi/raw   --data_corr ../data_test/umi/corrected   --size 100000   --outdir ../results_wf2

Outputs:

• UMI collision rates, duplication statistics
• Transcript-level QC

Purpose: Evaluate predicted isoforms against reference annotations (truth).

Command:

cd 3-Isoform_discovery
nextflow run main.nf --samplesheet ../GTF_samplesheet.csv   --reference_gtf ../data_test/refs/gencode.gtf   --outdir ../results_wf3

Outputs:

• Precision / recall / F1 per tool
• Structural error rates
• Visualization of isoform novelty types

Purpose: Comparison to a known ground truth dataset.

Command:

cd 4-Comparison_with_ground_truth
nextflow run main.nf --results_roots ../results_wf1,../results_wf3   --truth ../data_test/truth   --outdir ../results_wf4

Outputs:

• Gene & isoform benchmarking reports
• Clustering metrics (ARI, cLISI)
• Summary score tables

dataType: single-cell or spatial

WF1

WF2

WF3

WF4

• Always run using fixed Git tag releases (e.g., v1.0)
• Track Nextflow execution reports for provenance

If you use scKenver in your research, please cite:

Hamraoui A. et al.
A systematic benchmark of bioinformatics methods for single-cell and spatial RNA-seq Nanopore long-read data
BioRxiv, 2025. DOI: doi.org/10.1101/2025.07.21.665920.

Zenodo DOI for test data:

10.5281/zenodo.17632476

For issues or suggestions:
➡️ GitHub Issues: https://github.com/alihamraoui/scKenver/issues