metric of pairwise differences between two cells.

Inspired by MSKCC SPECTRUM paper. Bins are aligned between two cells and marked for if they have the same state Then segments of matching and non-matching runs of bins are found (filtering those smaller than a specified minimum). These segments are then re-split into 500kb bins and the difference becomes the number of matching bins divided by the number of considered bins.

Usage

pairwise_bin_difference(
  bin_df,
  targ_cells = c(),
  min_seg_length = 2500000,
  return_pairs_matrix = FALSE
)

Arguments

bin_df

a dataframe of read bins with states. Expected columns of: cell_id, chr, start, end, state

targ_cells

optional vector specifying which cells to compare to all other cells.

min_seg_length

double. This is the minium length of matching segment bins to use when measuring similarity.

Value

tibble of cell pairs and metrics about their differences.

Details

This function is slow, and the number of pairwise comparisons grows quickly. Dramatic speed improvements can be had by setting up a parallel plan for furrr like so:

future::plan(future::multisession, workers=N_CORES_YOU_WANT)

Example for 100 cells, which is 4950 pairs, this function will take 4 minutes with 4 cores.

These comparisons are unique pairs! So if 3 input cells: A, B, C, comparisons made are A-B, A-C, B-C. So one input cell will always be "missing" from the index column. See the vignette for how you can turn this around, but it can create a very large DF with redundant comparisons to rearrange to have each cell against all others.