Collapse Windows Based on Methylation Differences

This function collapses significant windows in a methylation dataset by merging contiguous regions that meet the specified criteria. Can only collapse windows once a differential modification analysis (calc_mod_diff()) has been called.

Usage

collapse_mod_windows(
  mod_db,
  table_name = "collapsed_windows",
  max_distance = 1000,
  sig_cutoff = 0.05,
  min_diff = 0.5
)

Arguments

mod_db

A DuckDB database connection object or path to the database.

table_name

Character. Name of the output table to store collapsed windows (default: "collapsed_windows").

max_distance

Numeric. The maximum allowable distance between consecutive significant windows for merging (default: 1000).

sig_cutoff

Numeric. The significance threshold for adjusted p-values (default: 0.05).

min_diff

Numeric. The minimum absolute methylation difference required for inclusion in the analysis (default: 0.5).

Value

This function does not return an object; it creates or replaces the `collapsed_windows` table in the database.

Details

The function performs the following steps:

• Filters the `mod_diff_windows` to retain only significant windows where `p_adjust <= sig_cutoff` and `ABS(meth_diff) >= min_diff`.

• Assigns a new region identifier based on proximity (`max_distance`) and the direction of methylation differences.

• Collapses regions by grouping contiguous windows, computing the average methylation difference (`avg_meth_diff`), and counting the number of merged windows.