SeATAC

SeATAC employs a conditional variational autoencoder to detect differential chromatin accessibility from ATAC-seq V-plots across the genome.

Key Features:

• Conditional variational autoencoder: Learns and interprets latent representations of ATAC-seq V-plots for chromatin accessibility analysis.
• Genome-wide differential detection: Identifies genomic regions with differential chromatin accessibility.
• Benchmarking performance: Outperforms MACS2 and NucleoATAC in six distinct chromatin accessibility analysis tasks.
• Quantitative site-level analysis: Detects reductions in chromatin accessibility at 20% to 30% of pioneer factor target sites.

Scientific Applications:

• Pioneer factor analysis: Applied to pioneer factor–induced datasets to study chromatin remodeling during cellular differentiation and reprogramming.
• Chromatin accessibility mapping: Used to detect and quantify differential chromatin accessibility from ATAC-seq data.
• Comparative method evaluation: Employed to benchmark peak-calling and nucleosome-detection tools such as MACS2 and NucleoATAC across multiple tasks.

Methodology:

Uses a conditional variational autoencoder to learn latent representations of ATAC-seq V-plots and compares performance against MACS2 and NucleoATAC across six benchmarking tasks.