Fig. 1: PACS (Probability model of Accessible Chromatin in Single Cells) framework.
From: Depth-corrected multi-factor dissection of chromatin accessibility for scATAC-seq data with PACS
a Upper panel: Illustration of the latent accessibility of cells. Multiple factors such as cell types, developmental stages, and spatial locations determine the chromatin structure and configurations within respective cell groups. These different chromatin structures result in the variable Tn5 insertion rates in the ATAC-seq experiments. The readout of ATAC assays is paired insertion counts (PIC), which are crude measures of latent insertion rates. Middle panel: Schematic of the read-capturing process in scATAC-seq. During PCR and sequencing, fragments in each single cell are partially captured, and after data processing, variable capturing probability should be accounted for in data modeling. Lower panel: Example downstream application of PACS, including compound hypothesis testing, cell type label transfer, and feature-level batch effect correction. b Example chromatin configurations where ignoring factors in Differential Accessible Regions (DARs) test may result in false interpretation.
