Table 1 Comparison of model design and capability among existing scATAC-seq testing tools
From: Depth-corrected multi-factor dissection of chromatin accessibility for scATAC-seq data with PACS
PACS | Seurat | ArchR | snapATAC | snapATAC2 | Contingency table test | ||
|---|---|---|---|---|---|---|---|
Method overview | Model type | missing corrected cumulative logistic regression (mcCLR) | Logistic regression | Wilcoxon rank sum test | edgeR Negative Binomial test | Logistic regression | Fisher’s exact test |
Sequencing depth adjustment | Probability model using data matrix | Total reads as a covariate in regression | Downsampling cells into matched depth | Normalizing in pseudobulk | Log(total reads) as a covariate in regression | ✗ | |
Model-based control of sampling imbalance | ✓ | ✗ | ✗ | ✓ | ✗ | ✗ | |
Model-based control of data sparsity | ✓ | ✗ | ✗ | ✗ | ✗ | ✗ | |
Multi-factor statistical tests | Multi-factor testing (built-in) | ✓ | ✓ | ✗ | ✗ | ✓ | ✗ |
Testing interaction between factors | ✓ | ✗ | ✗ | ✗ | ✗ | ✗ | |
Batch effect peak identification | ✓ | ✗ | ✗ | ✗ | ✗ | ✗ | |
Time series data analysis | ✓ | ✗ | ✗ | ✗ | ✗ | ✗ |
