paired-insertion-counting-strategy

paired-insertion-counting-strategy

Summary

A fragment quantification method in SnapATAC2 that counts chromatin accessibility at fixed genomic tiles by processing paired-end insertion coordinates from single-cell ATAC-seq fragment data. This strategy generates a cell-by-tile count matrix suitable for downstream analysis of chromatin accessibility patterns.

When to use

Apply this skill when you have loaded fragment data from single-cell ATAC-seq experiments into a backed AnnData object (with fragments stored in .obsm['fragment_paired'] or .obsm['fragment_single']) and need to quantify chromatin accessibility across fixed genomic intervals (tiles) rather than at predefined peaks or genes.

When NOT to use

• Fragment data is unavailable or not loaded into the AnnData object structure
• Analysis goal requires peak-level or gene-level quantification instead of genome-wide tiles (use pp.make_peak_matrix or pp.make_gene_matrix respectively)
• Fragments have not been processed or validated for quality (remove low-quality fragments or cell barcodes first via pp.filter_cells)

Inputs

• Backed AnnData object with fragment data in .obsm['fragment_paired'] or .obsm['fragment_single']
• Genomic interval specifications (tile width, typically 5 kb)

Outputs

• Sparse count matrix (n_cells × n_tiles) stored in .X or designated matrix slot
• Updated AnnData object with tile-based accessibility counts

How to apply

Invoke pp.add_tile_matrix with counting_strategy='paired_insertion' on a backed AnnData object containing fragment data. The function processes paired-end fragments by counting insertion events within fixed-width genomic tiles (typically 5 kb or user-specified width) across the entire genome. Each cell's fragment insertions are aggregated into bins, producing a sparse count matrix where rows are cells and columns are genomic tiles. After execution, verify the output matrix shape (n_obs × n_vars) matches the number of cells and expected tile coordinates, and confirm that count values are non-zero and distributed across cells and tiles with expected sparsity patterns typical of chromatin accessibility data.