Search code examples
rperformancedata.tablegenomicranges

How to remove overlapping sequences from two datatables?


I have two data.tables that provide sequence coordinates across different chromosomes (categories). For example:

library(data.table)
dt1 <- data.table(chromosome = c("1", "1", "1", "1", "X"),
                  start = c(1, 50, 110, 150, 110),
                  end = c(11, 100, 121, 200, 200))
dt2 <- data.table(chromosome = c("1", "1", "X"),
                  start = c(12, 60, 50),
                  end = c(20, 115, 80))

I need to create a third data.table that provides coordinates for sequences that contain all the integers in dt1 that do not overlap with any integers from the sequences in dt2. For example:

dt3 <- data.table(chromosome = c("1", "1", "1", "1", "X"),
                  start = c(1, 50, 116, 150, 110),
                  end = c(11, 59, 121, 200, 200))

The data.tables I need to run this on are very large and therefore I need to maximise performance. I have tried using the foverlaps() function but to no avail. Any help would be greatly appreciated!


Solution

  • You can start with something like this from foverlaps

    setkey(dt2,chromosome,start,end)
    ds = foverlaps(dt1,dt2,  type="any")
    ds[,.(chromosome, 
          start = fcase(is.na(start) | i.start <= start,i.start,
                        i.end >= end, end + 1),
          end = fcase(is.na(end) | i.end >= end, i.end,
                      i.start <= start, start - 1)
          )]
    #   chromosome start   end
    #       <char> <num> <num>
    #1:          1     1    11
    #2:          1    50    59
    #3:          1   116   121
    #4:          1   150   200
    #5:          X   110   200