Question

I want to split a data frame like this

chr.pos nt.pos  CNV
1   74355   0
1   431565  0
1   675207  0
1   783605  1
1   888149  1
1   991311  1
1   1089305 1
1   1177669 1
1   1279886 0
1   1406311 0
1   1491385 0
1   1579761 0
2   1670488 1
2   1758800 1
2   1834256 0
2   1902924 1
2   1978088 1
2   2063124 0 

The point is to get a list of intervals where the chr are the same and CNV=1 column, but taking into account the 0 inervals between them

[[1]]

   1    783605  1
   1    888149  1
   1    991311  1
   1    1089305 1
   1    1177669 1
[[2]]
   2    1670488 1
   2    1758800 1
[[3]]
   2    1902924 1
   2    1978088 1

Any ideas?

Answer 1

You can use rle to create a variable to use in split

  # create a group identifier
  DF$GRP <- with(rle(DF$CNV), rep(seq_along(lengths),lengths))


  # split a subset of DF which contains only CNV==1
  split(DF[DF$CNV==1,],DF[DF$CNV==1,'GRP'] )


$`2`
  chr.pos  nt.pos CNV GRP
4       1  783605   1   2
5       1  888149   1   2
6       1  991311   1   2
7       1 1089305   1   2
8       1 1177669   1   2

$`4`
   chr.pos  nt.pos CNV GRP
13       2 1670488   1   4
14       2 1758800   1   4

$`6`
   chr.pos  nt.pos CNV GRP
16       2 1902924   1   6
17       2 1978088   1   6