PeteHaitch

Keybase proof

I hereby claim:

• I am petehaitch on github.
• I am peterhickey (https://keybase.io/peterhickey) on keybase.
• I have a public key ASCF4DS1IJgSk7plC9JD2iCp-ey_j53id8_e82va8CmkLQo

To claim this, I am signing this object:

PeteHaitch

library(profmem)
library(Matrix)
dgt <- readMM("~/sparse.mtx.gz")
dgc <- as(dgt, "dgCMatrix")

colCountsEqualZero <- function(x) {
  if (length(x@x) / length(x) > 0.5) {
    # If majority of data are non-zero
    Matrix::colSums(x == 0)
  } else {

PeteHaitch

library(profmem)
library(Matrix)
dgt <- readMM("~/sparse.mtx.gz")
dgc <- as(dgt, "dgCMatrix")
threshold <- 5

# Total memory allocation (bytes)
total(profmem(Matrix::colSums(dgt > threshold)))
#> [1] 935422632
total(profmem(Matrix::colSums(dgc > threshold)))

PeteHaitch

library(Matrix)
library(pryr)
library(profmem)
library(matrixStats)
library(microbenchmark)

nrow <- 20000
ncol <- 600
threshold <- 0

PeteHaitch

library(GenomicRanges)

# A fast check that all(x == y) for GenomicRanges objects. 
.all.equal.GenomicRanges <- function(x, y) {
  seqinfo <- merge(seqinfo(x), seqinfo(y))
  seqlevels <- seqlevels(seqinfo)
  if (any(diff(match(seqlevels(y), seqlevels)) < 0L)) {
    stop("the 2 objects to compare have seqlevels in incompatible orders")
  }
  ok <- all(identical(seqnames(x), seqnames(y)),

PeteHaitch

# Pseudocode
isUnsortedListOfAtomics <- function(..., na.rm = FALSE, strictly = FALSE) {
  args <- list(...)
  m <- length(args)
  n <- length(args[[1]])

  for (i in (seq_along(args[[1]]) - 1)) {
    for (j in seq_len(m)) {
      # All vectors except the last are checked in the same way 
      if (j < m) {

PeteHaitch

# (1) Using existing repo
git clone [email protected]:PeteHaitch/GenomicTuples.git
cd GenomicTuples
curl -O https://raw.githubusercontent.com/Bioconductor/mirror/master/update_remotes.sh
bash update_remotes.sh
# Bump Version and Date in DESCRIPTION (v1.5.2)
# NOTE: Version isn't up-to-date with SVN; why?
git add DESCRIPTION
git commit -m "Bump version number"

PeteHaitch

Combining SummarizedExperiment objects

Peter Hickey
20 October 2015

Motivation

I often find myself with multiple SE objects (I'm using SE as a shorthand for the SummarizedExperiment0 and RangedSummarizedExeriment classes), each with potentially non-distinct samples and potentially non-overlapping features/ranges. Currently, it is difficult to combine these objects; rbind() can only combine objects with the same samples but distinct features/ranges and cbind() can only combine objects with the same features/ranges but distinct samples. I think it would be useful to have a "combine" method for SE objects that handles the most general situation where each object has potentially non-distinct samples and potentially non-overlapping features/ranges.

PeteHaitch

library(GenomicRanges)
library(microbenchmark)

nrows <- 2000000; ncols <- 6
counts <- matrix(runif(nrows * ncols, 1, 1e4), nrows)
rowData <- GRanges(rep(c("chr1", "chr2"), c(0.25 * nrows, 0.75 * nrows)),
                   IRanges(floor(runif(nrows, 1e5, 1e6)), width=100),
                   strand=sample(c("+", "-"), nrows, TRUE))
colData <- DataFrame(Treatment=rep(c("ChIP", "Input"), 3),
                     row.names=LETTERS[1:6])

PeteHaitch

## Necessary packages (BioC-devel)
library(GenomicRanges)
library(S4Vectors)

## Class A using a DataFrameOrNULL in internalPos slot
setClassUnion(name = "DataFrameOrNULL", members = c("DataFrame", "NULL"))

setClass("A",
         contains = "GRanges",
         representation(