Find the best overlap between GRanges

Find the best overlap between ranges

bestOverlap(x, y, ...)

# S4 method for class 'GRanges,GRanges'
bestOverlap(
  x,
  y,
  var = NULL,
  ignore.strand = FALSE,
  missing = NA_character_,
  min_prop = 0.01,
  ...
)

# S4 method for class 'GRanges,GRangesList'
bestOverlap(
  x,
  y,
  ignore.strand = FALSE,
  missing = NA_character_,
  min_prop = 0.01,
  ...
)

Arguments

x

a GRanges object

y

a named GRangesList or GRanges object with mcol as reference category

...

Not used

var

The variable to use as the category. Not required if y is a GRangesList

ignore.strand

logical(1) Passed to findOverlaps

missing

Value to assign to ranges with no overlap

min_prop

Threshold below which overlaps are discarded

Value

Character vector the same length as the supplied GRanges object

Details

This finds the category in the subject GRanges (y) which has the best overlap with the query GRanges (x). The aim is to produce a character vector for best classifying the query GRanges using an external set of features (e.g. promoters, enhancers etc). If the subject (y) is a GRanges object, the values in the specified column will be used as the category. If the subject (y) is a GRangesList, the names of the list will be used to provide the best match

Examples

gr <- GRanges("chr1:1-10")
gr_cat <- GRanges(c("chr1:2-10", "chr1:5-10"))
gr_cat$category <- c("a", "b")
propOverlap(gr, gr_cat)
#> [1] 0.9
bestOverlap(gr, gr_cat, var = "category")
#> [1] "a"

grl <- splitAsList(gr_cat, gr_cat$category)
lapply(grl, function(x) propOverlap(gr, x))
#> $a
#> [1] 0.9
#> 
#> $b
#> [1] 0.6
#> 
bestOverlap(gr, grl)
#> [1] "a"