binstats

Given GRanges of binned eg read depth data with field $cn, crosses nodes in graph and annotates graph nodes with 'cn' and 'weight'. Done upstream of balance.

If "by" field(s) specified and these fields exist in both bins and graph nodes, then will use these in the overlaps query

If field, purity, and ploidy provided then will also transform read depth data in bin column "field" using purity and ploidy to generate

binstats(
  gg,
  bins,
  by = NULL,
  field = NULL,
  purity = gg$meta$purity,
  ploidy = gg$meta$ploidy,
  loess = TRUE,
  min.bins = 3,
  verbose = TRUE,
  min.var = 0.1,
  lp = FALSE
)

Arguments

gg

gGraph

bins

GRanges with field $cn or field field

by

optional character vector specifying metadata field(s) shared by gg$nodes and bins to which bin / node overlaps will be limited to

field

character vector field of bins to convert to (NULL)

purity

purity parameter either specified together with field or embedded in gg$meta, must be specified if field is not NULL

ploidy

ploidy parameter either specified together with field or embedded in gg$meta, must be specified if field is not NULL

loess

logical flag whether to smooth / fit variance using loess (FALSE)

min.bins

minimum number of bins to use for intra segment variance computation (3)

min.var

minimal allowable per segment bin variance, which will ignore segments with very low variance due to all 0 or other reasons (0.1)

lp

(logical) return weights consistent with LP optimization

Value

gGraph whose nodes are annotated with $cn and $weight field

Author

Marcin Imielinski