On a twitter exchange I showed A. Quinlan how to use epivizr to load data from a custom bed file. Here is the code I gave him:
library(epivizr)
library(rtracklayer)
# download example bed file
download.file("https://raw.githubusercontent.com/arq5x/bedtools/master/data/aluY.hg19.bed.gz", destfile="test.bed.gz", method="curl")
# start UI
mgr <- startEpiviz(workspace="mi9NojjqT1l")
# import bed file
gr <- import(BEDFile("test.bed.gz"))
# drop data from unplaced contigs
gr <- keepSeqlevels(gr, paste0("chr",c(1:22,"X","Y")))
# add track with bed file data
dev <- mgr$addDevice(gr, "example bed")
# finish up
mgr$stopServer()
Also at https://gist.github.com/hcorrada/f930fa0092f1100f1d37.
The idea behind the epivizr BioC package is that it can use that infrastructure to import a lot of data formats into GenomicRanges-like objects one can manipulate (say, filter or transform), and have interactive visualization that reflects those manipulations immediately. However, it’s a little cumbersome for the use-case of where you have data on a BED file that you don’t need to manipulate, but just explore visually.
An option we’d like to get started with to support this contextual-data use-case is to write small programs that would use, say bedtools for example, to implement the epiviz Data Provider API and serve data directly from a bed file.