run_scc_scan: Compute the SCC scan statistic for a sequence of window...

In hillarykoch/locdiffr: analysis of LOCal DIFFerences in chromatin aRchitecture

Description

This function implements the SCC scan and transforms the scan to Z-scores. It can be paralllized across a grid of window sizes.

Usage

run_scc_scan(infiles1, infiles2, outpath, resolution, winsizes = seq(10, 100, by = 10), min_count = 1, proportion_below_min_count = 0.25, parallel = FALSE, ncores = 10, offset = TRUE, return = FALSE)

Arguments

infiles1	List or vector of paths to data files from the first condition.
infiles2	List or vector of paths to data files from the second condition.
outpath	Path to where the output will be written. The output file should be given the .rds file extension.
resolution	Resolution of the data being analysis, e.g., 50000 for 50 kilobase resolution.
winsizes	A vector of window sizes to use for the scan, in bins.
min_count	A desired minimum number of counts within each bin. See details.
proportion_below_min_count	If the proportion of bins within a window with counts less than "min_count" is greater than this value, the window is filtered. See details.
parallel	Logical. Should the function be parallelized?
ncores	If parallel=TRUE, how many cores to parallelize over.
offset	Should an offset of 1 be used? Use FALSE if the input files are one-indexed, but TRUE if the input files are zero-indexed.
return	Logical. Should the output also be returned by the function?

Details

This function writes the output to the provided outpath in an rds file.

Windows with insufficient reads can be filtered from the analysis using "min_count" and "proportion_below_min_count" options. The default sets min_count = 1 and proportion_below_min_count = 0.25. This means that if more than 25% of the bins within a window have less than 1 measured interaction, the window is filtered from downstream analyses. To do no filtering, set proportion_below_min_count=1.

Value

If return = TRUE, this function returns a list of length winsizes which contains the SCC scan for each window size.

Examples

# Define paths to data from 2 conditions
all_data <- list.files(system.file("extdata", package = "sgp"))
ko_data <-
    system.file("extdata", all_data[grepl(pattern = "KO", x = all_data)], package = "sgp")
parent_data <-
    system.file("extdata", all_data[grepl(pattern = "parent", x = all_data)], package = "sgp")

# Compute the SCC scan statistics
run_scc_scan(
    infiles1 = ko_data,
    infiles2 = parent_data,
    outpath = "inst/extdata/output/z.rds",
    resolution = 40000,
    winsizes = c(10, 25),
    parallel = FALSE,
    offset = TRUE
)

hillarykoch/locdiffr documentation built on April 25, 2021, 6:26 p.m.