In charles-plessy/CAGEr: Analysis of CAGE (Cap Analysis of Gene Expression) sequencing data for precise mapping of transcription start sites and promoterome mining

Originally, the .plotReverseCumulative function was producing sorted abundances using the data.table package, as in the function f.dt declared below. After benchmarking it with alternativesk, I switched to the approach implemented in f.Rle because the data fed to .plotReverseCumulative is often already in Rle format.

options(width=120)
knitr::opts_chunk$set(cache  = TRUE, cache.lazy = FALSE)
knitr::opts_knit$set(verbose = TRUE)

set.seed(1)
pois <- rpois(1e7, 0.3)
nbin <- rnbinom(1e7, mu = 2, size = .1)

f.df   <- function(x) as.data.frame(table(x))

f.dt   <- function(x) {
  v <- data.table::data.table(num = 1, nr_tags = x)
  v <- v[, sum(num), by = nr_tags]
  data.table::setkeyv(v, "nr_tags")
  as.data.frame(v)
}

f.Rle  <- function(x) {
  x <- S4Vectors::Rle(sort(x))
  data.frame(S4Vectors::runValue(x), S4Vectors::runLength(x))
}

f.rle  <- function(x) {
  x <- rle(sort(x))
  data.frame(x$values, x$lengths)
}

f.aggr <- function(x) {
  aggregate(x, by = list(x), FUN = length)
}

microbenchmark::microbenchmark(f.df(pois), f.dt(pois), f.Rle(pois), f.rle(pois), f.aggr(pois), times = 10)

microbenchmark::microbenchmark(f.df(nbin), f.dt(nbin), f.Rle(nbin), f.rle(nbin), f.aggr(nbin), times = 10)

charles-plessy/CAGEr documentation built on Aug. 2, 2024, 4:35 p.m.