SCRIMPy is a Nextflow pipeline that processes single-cell ATAC-seq fragment data to infer DNA replication states of individual cells. The main steps are:
- Filter ATAC fragments using provided list of cell barcodes and retain only autosomal fragments for each cell.
- Infer replication timing from fragment coverage if replication timing data is not provided.
- Assign replication timing rank to each fragment.
- Identify overlapping fragments.
- Use obtained replication and overlap information of fragments to compute per-cell metrics file.
- Nextflow (>=23.04.2)
- Java 11 or later
- Python (>= 3.11.3)
- BEDTools (>= 2.31.0)
git clone https://github.com/TatevikJ/scrimpy.git
cd scrimpynextflow run main.nf -c nextflow.config Before running the pipeline, prepare the .config file with the correct input parameters.
| Parameter | Description |
|---|---|
input_paths_csv |
Path to sample sheet (CSV format) |
outdir |
Output directory |
The pipeline requires a CSV sample sheet with the following columns:
sample,fragments,barcodes
sample1,/path/to/sample1/atac_fragments.tsv.gz,/path/to/sample1/barcode_list.tsv
sample2,/path/to/sample2/atac_fragments.tsv.gz,/path/to/sample2/barcode_list.tsv| Parameter | Description | Default |
|---|---|---|
rt_tsv |
Path to replication timing data (false if using RT inference) |
None (required if rt_chrom_sizes and rt_fragments are not provided) |
rt_chrom_sizes |
Path to chrom.sizes file for generating genomic bins for RT inference |
None (required if rt_tsv is false) |
rt_fragments |
Path to a fragment file (atac_fragments.tsv.gz) used for RT inference |
None (required if rt_tsv is false) |
mem_per_cpu |
Memory allocated per CPU | 40.GB |
cpus |
Number of CPUs to use per task | 6 |
outdir/
├── reports/ # Pipeline execution reports
├── sample1/ # Sample 1 analysis results
│ └── metrics_per_cell.csv # Per-cell metrics for sample 1
│ └── ...
└── sample2/ # Sample 2 analysis results
└── metrics_per_cell.csv # Per-cell metrics for sample 1
└── ...