Find simple deletions and rigmas

Calls simple deletions (del) and rigma, which are "rifts" or clusters of overlapping deletions

del(
  gg,
  fdr.thresh = 0.5,
  tile.width = 1e+06,
  cn.thresh = 2,
  jcn.thresh = 1,
  min.count = 2,
  mark = FALSE,
  return.fish = FALSE,
  mark.col = "purple",
  min.width = 10000,
  max.width.flank = 10000,
  max.width = 1e+07
)

Arguments

gg

gGraph with $cn field annotated on nodes and edges

fdr.thresh

False Discovery Rate threshold. Default: 0.5

tile.width

bin width to use when computing deletion clustering. Default: 1e6

cn.thresh

node copy number threshold to call a deletion in ploidy units. Default: 2

jcn.thresh

edge copy number threshold to call a deletion in ploidy units. Default: 1

min.count

minimum number of overlapping deletions to constitute a rigma, i ncludes the tile.width. Default: 2

mark

color deletion events. Default: False

return.fish

Parameter to return fishook::Fish output. Default: False

mark.col

color of event. Default: purple

min.width

min width of deletions to consider for rigma. Default: 1e4

max.width.flank

max width flank each side of class dup-like to consider. Default: 1e4

max.width

max width of deletions to consider for rigma. Default: 1e7

Value

gGraph with nodes and edges annotated with $del and $rigma metadata field, and data.tables $meta$rigma and $meta$set with event level statistics.

If return.fish = TRUE, returns FishHook::Fish() output.

Details

Deletions are defined as having low junction copy number and connect two nodes of low junction copy number (with cn and jcn thresholds provided as parameters). Simple deletions have no other overlapping junctions in their shadow. Rigma have a min.count and are also outliers (<fdr.thresh) in a negative binomial model that incorporates the regional (non-DEL) junction count in tile.width genomic bins (set to 1 Mbp by default).

Note: Not all DEL-like junctions will be called a del or rigma.

More details on how to run this function and examples: Deletions & Rigma