RleDataFrame-views: Calculate summary statistics on views of an RleDataFrame

Description

These methods mirror the viewMeans type functions from IRanges for SimpleRleList. They differ in that they work on an RleDataFrame and an IRanges directly and also have a simplify argument. This works out to be faster (compute-wise) and also convenient.

Still, an RleDataFrame inherits from SimpleRleList, so all of the views functions will work.

Usage

rangeSums(x, bounds, na.rm=FALSE, simplify=TRUE)
rangeMeans(x, bounds, na.rm=FALSE, simplify=TRUE, ...)
rangeMins(x, bounds, na.rm=FALSE, simplify=TRUE)
rangeMaxs(x, bounds, na.rm=FALSE, simplify=TRUE)
rangeWhichMins(x, bounds, na.rm=FALSE, simplify=TRUE)
rangeWhichMaxs(x, bounds, na.rm=FALSE, simplify=TRUE)

Arguments

RleDataFrame

bounds

Matrix with two columns or IRanges representing ranges of rows of x to process. If bounds is a matrix, an IRanges is constructed assuming the first two columns represent the start and end of the ranges. The names for the IRanges is taken from the rownames of the matrix. Such a matrix can constructed with boundingIndicesByChr and is the preferred input.

na.rm

Scalar logical. Ignore NAs in calculations?

simplify

Scalar logical. Simplify result? If TRUE, the return value will be a vector or matrix. For a single view, a vector will be returned. Otherwise a matrix with one row per view and one column per column of x will be returned. If FALSE, the return value will be a list of length ncol(x) of vectors of length nrow(bounds).

...

Additional arguments for other methods.

Value

With simplify == TRUE, a vector for single view or a matrix otherwise. When simplify == FALSE, a list of vectors length ncol(x) where each element is of length nrows(bounds).

Details

The "range" name prefixes here serve to differentiate these functions from the "view" functions. This may change. I will be asking the IRanges team to add "..." and "simplify" to the "view" methods so that I can just make additional methods for RleDataFrame.

Examples

Run this code

df = RleDataFrame(list(a=Rle(1:5, rep(2, 5))), b=Rle(1:5, rep(2, 5)),
row.names=LETTERS[1:10])
mat = matrix(c(1,4,3,5),ncol=2,dimnames=list(c("Gene1","Gene2"),c("start","end")))
bounds = IRanges(start=c(1, 4), end=c(3, 5), names=c("Gene1","Gene2"))

rangeMeans(df,bounds,simplify=FALSE)
rangeMeans(df,bounds,simplify=TRUE)
rangeMeans(df,mat,simplify=TRUE)

rangeMeans(df,bounds)
rangeSums(df,bounds)
rangeMins(df,bounds)
rangeMaxs(df,bounds)
rangeWhichMins(df,bounds)
rangeWhichMaxs(df,bounds)

# RleDataFrame isa SimpleRleList, so all the IRanges view* methods work too:
v = RleViewsList( lapply( df, Views, start=bounds ) )
viewMeans(v)