Filter an edgelist by downsampling read counts

Filter an edgelist (for a subset of components) in a PXL file by downsampling read counts using predefined fractions. The probability of a read staying in the edgelist is \(1 - (1 - frac) ^ read_count\).

Usage

downsample_to_parquet(pxl_file, outdir, components, fracs = seq(0.1, 0.9, 0.1))

Arguments

pxl_file

Path to the PXL file.

outdir

Directory to write the output parquet files.

components

A character vector of component names to keep.

fracs

A numeric vector of fractions to downsample by.

Value

A tibble with paths to the written parquet files and the fractions.

Details

The filtered edgelist(s) are written to parquet files in outdir.