simATAC: a single-cell ATAC-seq simulation framework.

Area of Science:

• Genomics
• Computational Biology
• Epigenetics

Background:

• Single-cell assay for transposase-accessible chromatin sequencing (scATAC-seq) is a powerful technique for dissecting cellular heterogeneity.
• Developing robust computational pipelines for scATAC-seq data analysis requires reproducible and well-characterized datasets for benchmarking.
• Existing datasets often lack the necessary characteristics for comprehensive pipeline evaluation.

Purpose of the Study:

• To introduce simATAC, an R package designed to generate synthetic scATAC-seq count matrices.
• To provide a tool that produces in silico scATAC-seq data closely mimicking real datasets in key features like library size, sparsity, and accessibility signals.
• To facilitate the rigorous assessment and comparison of various scATAC-seq analytical pipelines.

Main Methods:

• The simATAC framework utilizes statistical models derived from analyzing a large cohort of 90 real scATAC-seq cell groups.
• It generates count matrices that accurately reflect the properties of experimental scATAC-seq data.
• The package allows for the creation of simulated datasets with known cell labels.

Main Results:

• simATAC successfully generates synthetic scATAC-seq data with high fidelity to real experimental data.
• The simulated datasets exhibit comparable library sizes, sparsity patterns, and chromatin accessibility signal distributions.
• The framework enables the generation of controlled datasets for systematic evaluation of downstream analysis tools.

Conclusions:

• simATAC offers a robust and systematic approach for creating in silico scATAC-seq samples.
• This tool is essential for benchmarking and validating computational methods in single-cell epigenomics.
• The availability of simATAC will accelerate the development and improvement of scATAC-seq data analysis pipelines.