svtk bedcluster doesn't cluster events with identical coordinates

[mgonzalezporta]

E.g. for this input:

$ cat input.bed
chr10   105297001       105303000       chr10:105297001-105303000       WHB3855 deletion
chr10   105297001       105302000       chr10:105297001-105302000       WHB3873 deletion
chr10   105297001       105303000       chr10:105297001-105303000       WHB3880 deletion
chr10   105297001       105303000       chr10:105297001-105303000       WHB3882 deletion
chr10   105297001       105303000       chr10:105297001-105303000       WHB3884 deletion
chr10   105297001       105302000       chr10:105297001-105302000       WHB3904 deletion
chr10   105297001       105303000       chr10:105297001-105303000       WHB3934 deletion
chr10   105297001       105303000       chr10:105297001-105303000       WHB3939 deletion
chr10   105297001       105303000       chr10:105297001-105303000       WHB3961 deletion

Running svtk bedcluster gives the following output:

$ svtk bedcluster input.bed output.bed --merge-coordinates --prefix bedcluster --frac 0.8

$ cat output.bed
#chrom  start   end     name    svtype  sample  call_name       vaf     vac     pre_rmsstd      post_rmsstd
chr10   105297001       105303000       bedcluster_0    deletion        WHB3855 chr10:105297001-105303000       0.111   1       0.000   0.000
chr10   105297001       105302000       bedcluster_1    deletion        WHB3873 chr10:105297001-105302000       0.111   1       0.000   0.000
chr10   105297001       105303000       bedcluster_2    deletion        WHB3880 chr10:105297001-105303000       0.111   1       0.000   0.000
chr10   105297001       105303000       bedcluster_3    deletion        WHB3882 chr10:105297001-105303000       0.111   1       0.000   0.000
chr10   105297001       105303000       bedcluster_4    deletion        WHB3884 chr10:105297001-105303000       0.111   1       0.000   0.000
chr10   105297001       105302500       bedcluster_5    deletion        WHB3904 chr10:105297001-105302000       0.222   2       500.000 500.000
chr10   105297001       105302500       bedcluster_5    deletion        WHB3961 chr10:105297001-105303000       0.222   2       500.000 500.000
chr10   105297001       105303000       bedcluster_6    deletion        WHB3934 chr10:105297001-105303000       0.111   1       0.000   0.000
chr10   105297001       105303000       bedcluster_7    deletion        WHB3939 chr10:105297001-105303000       0.111   1       0.000   0.000

There's actually 6 calls that share the same coordinates, but they are reported with different bedcluster ids:

$ cat output.bed | cut -f 1-3 | grep -v '^#' | sort | uniq -c | sort -r
      6 chr10   105297001       105303000
      2 chr10   105297001       105302500
      1 chr10   105297001       105302000

A suggestion: ensure that events with identical coordinates have the same bedcluster id in the clustered output bed.

Thanks!