viewGRangesWinSummary_dt: Summarizes signal in bins. The same number of bins per...

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

Summarizes signal in bins. The same number of bins per region in qgr is used and widths can vary in qgr, in contrast to viewGRangesWinSample_dt where width must be constant across regions.

Description

This function is most appropriate where features are expected to vary greatly in size and feature boundaries are important, ie. gene bodies, enhancers or TADs.

Usage

viewGRangesWinSummary_dt(
  score_gr,
  qgr,
  n_tiles = 100,
  attrib_var = "score",
  attrib_type = NULL,
  fill_value = 0,
  anchor = c("center", "center_unstranded", "left", "left_unstranded")[1],
  summary_FUN = stats::weighted.mean
)

Arguments

score_gr	GRanges with a "score" metadata column.
qgr	regions to view by window.
n_tiles	numeric >= 1, the number of tiles to use for every region in qgr.
attrib_var	character name of attribute to pull data from. Default is "score", compatible with with bigWigs or bam coverage.
attrib_type	one of NULL, qualitative or quantitative. If NULL will attempt to guess by casting attrib_var attribute to character or factor. Default is NULL.
fill_value	numeric or character value to use where queried regions are empty. Default is 0 and appropriate for both calculated coverage and bedgraph/bigwig like files. Will automatically switch to "MISSING" if data is guessed to be qualitative.
anchor	character. controls how x value is derived from position for each region in qgr. 0 may be the left side or center. If not unstranded, x coordinates are flipped for (-) strand. One of c("center", "center_unstranded", "left", "left_unstranded"). Default is "center".
summary_FUN	function. used to aggregate score by tile. must accept x=score and w=width numeric vectors as only arguments. default is weighted.mean. limma::weighted.median is a good alternative.

Details

Columns in output data.table are: standard GRanges columns: seqnames, start, end, width, strand id - matched to names(score_gr). if names(score_gr) is missing, added as seq_along(score_gr). y - value of score from score_gr x - relative bp position

Value

data.table that is GRanges compatible

Examples

data(CTCF_in_10a_overlaps_gr)
bam_file = system.file("extdata/test.bam",
    package = "seqsetvis")
qgr = CTCF_in_10a_overlaps_gr[1:5]
# unlike viewGRangesWinSample_dt, width is not fixed
# qgr = GenomicRanges::resize(qgr, width = 500, fix = "center")
bam_gr = seqsetvis:::fetchBam(bam_file, qgr)
bam_dt = viewGRangesWinSummary_dt(bam_gr, qgr, 50)

if(Sys.info()['sysname'] != "Windows"){
    bw_file = system.file("extdata/MCF10A_CTCF_FE_random100.bw",
        package = "seqsetvis")
    bw_gr = rtracklayer::import.bw(bw_file, which = qgr)
    bw_dt = viewGRangesWinSummary_dt(bw_gr, qgr, 50)
}

jrboyd/seqsetvis documentation built on Dec. 10, 2024, 11:23 a.m.