ssvFetchSignal: signal loading framework

In jrboyd/seqsetvis: Set Based Visualizations for Next-Gen Sequencing Data

signal loading framework

Description

Does nothing unless load_signal is overridden to carry out reading data from file_paths (likely via the appropriate ssvFetch* function, ie. ssvFetchBigwig or ssvFetchBam

Usage

ssvFetchSignal(
  file_paths,
  qgr,
  unique_names = NULL,
  names_variable = "sample",
  file_attribs = NULL,
  win_size = 50,
  win_method = c("sample", "summary")[1],
  return_data.table = FALSE,
  load_signal = function(f, nam, qgr) {
     warning("nothing happened, ",
    "supply a function to", "load_signal parameter.")
 },
  n_cores = getOption("mc.cores", 1),
  n_region_splits = 1,
  force_skip_centerFix = FALSE
)

Arguments

file_paths	character vector of file_paths to load from. Alternatively, file_paths can be a data.frame or data.table whose first column is a character vector of paths and additial columns will be used as metadata.
qgr	GRanges of intervals to return from each file
unique_names	unique file ids for each file in file_paths. Default is names of file_paths vector
names_variable	character, variable name for column containing unique_names entries. Default is "sample"
file_attribs	optional data.frame/data.table with one row per item in file paths. Each column will be a variable added to final tidy output.
win_size	numeric/integer window size resolution to load signal at. Default is 50.
win_method	character. one of c("sample", "summary"). Determines if viewGRangesWinSample_dt or viewGRangesWinSummary_dt is used to represent each region in qgr.
return_data.table	logical. If TRUE data.table is returned instead of GRanges, the default.
load_signal	function taking f, nam, and qgr arguments. f is from file_paths, nam is from unique_names, and qgr is qgr. See details.
n_cores	integer number of cores to use. Uses mc.cores option if not supplied.
n_region_splits	integer number of splits to apply to qgr. The query GRanges will be split into this many roughly equal parts for increased parallelization. Default is 1, no split.
force_skip_centerFix	boolean, if TRUE all query ranges will be used "as is". This is already the case by default if win_method == "summary" but may have applications where win_method == "sample".

Details

load_signal is passed f, nam, and qgr and is executed in the environment where load_signal is defined. See ssvFetchBigwig and ssvFetchBam for examples.

Value

A GRanges with values read from file_paths at intervals of win_size. Originating file is coded by unique_names and assigned to column of name names_variable. Output is data.table is return_data.table is TRUE.

Examples

library(GenomicRanges)
bam_f = system.file("extdata/test.bam",
    package = "seqsetvis", mustWork = TRUE)
bam_files = c("a" = bam_f, "b" = bam_f)
qgr = CTCF_in_10a_overlaps_gr[1:2]
qgr = resize(qgr, 500, "center")
load_bam = function(f, nam, qgr) {
    message("loading ", f, " ...")
    dt = seqsetvis:::ssvFetchBam.single(bam_f = f,
                      qgr = qgr,
                      win_size = 50,
                      fragLen = NULL,
                      target_strand = "*",
                      return_data.table = TRUE)

    data.table::set(dt, j = "sample", value = nam)
    message("finished loading ", nam, ".")
    dt
}
ssvFetchSignal(bam_files, qgr, load_signal = load_bam)

jrboyd/seqsetvis documentation built on March 17, 2024, 3:14 p.m.