Problem using GenomicRanges package

Question

[tristan.verhoeff]

Hello,

I am trying to run the GRanges package after reading in a .csv file in Rstudio, I believe it is looking at methylation positions across chromosomes. The code was provided to me and is as follows:

islandHMM <- read.csv(paste0("C:/Users/Tristanv/Documents/RStudio/model-based-cpg-islands-hg19-chr17.txt"), header=FALSE, sep="\t", stringsAsFactors=FALSE)
islandData <- GRanges(seqnames=Rle(islandHMM[,1]), ranges=IRanges(start=islandHMM[,2], end=islandHMM[,3]), strand=Rle(strand(rep("*",nrow(islandHMM)))))

The csv file reads correctly and looks like the attached (first six rows shown).

The second part of the code concerning GRanges gives an error:

Error in .normargSEW0(start, "start") :    'start' must be a numeric vector (or NULL)

I am not sure how I would go about fixing this, I am fairly new to R.

Thanks in advance for and help.

[ATpoint]

Please provide output of dput(head(islandHMM, 20)) so one can have something to reproduce/provide code on.

[tristan.verhoeff]

That gives the following output (this is after the first entry in columns V2 to V7 has been turned into a double, so the first in each column is now NA):

structure(list(V1 = c("chr", "chr10", "chr10", "chr10", "chr10", "chr10", "chr10", "chr10", "chr10", "chr10", "chr10", "chr10", "chr10", "chr10", "chr10", "chr10", "chr10", "chr10", "chr10", "chr10"), V2 = c(NA, 93098, 94002, 94527, 119652, 122133, 180265, 180865, 242994, 313778, 319183, 321809, 323315, 323625, 324003, 327172, 334493, 346533, 348952, 357295), V3 = c(NA, 93818, 94165, 95302, 120193, 122621, 180720, 182549, 243152, 313905, 319290, 321998, 323465, 323770, 324088, 327614, 334801, 346829, 349176, 357450), V4 = c(NA, 721, 164, 776, 542, 489, 456, 1685, 159, 128, 108, 190, 151, 146, 86, 443, 309, 297, 225, 156), V5 = c(NA, 32, 12, 65, 53, 51, 32, 230, 10, 6, 8, 10, 11, 11, 7, 24, 14, 30, 16, 12), V6 = c(NA, 403, 97, 538, 369, 339, 256, 1263, 74, 64, 68, 113, 92, 87, 60, 248, 171, 210, 105, 106), V7 = c(NA, 0.559, 0.591, 0.693, 0.681, 0.693, 0.561, 0.75, 0.465, 0.5, 0.63, 0.595, 0.609, 0.596, 0.698, 0.56, 0.553, 0.707, 0.467, 0.679), V8 = c("obsExp", "0.572", "0.841", "0.702", "0.866", "0.88", "0.893", "0.984", "1.193", "0.769", "0.75", "0.596", "0.791", "0.991", "0.672", "0.692", "0.594", "0.833", "1.326", "0.667" )), row.names = c(NA, 20L), class = "data.frame")

Answer 1

[bernatgel]

Hi Tristan,

The problem you have is that your tell read.csv that your file has no header (header=FALSE) but it has a header, and so the first data line in your data.frame is all characters and forces all columns to be characters. To solve that, simply change header=FALSE to header=TRUE and everything should work.

Another option would be to use toGRanges from the package regioneR

library(regioneR)
islandHMM <- toGRanges(""C:/Users/Tristanv/Documents/RStudio/model-based-cpg-islands-hg19-chr17.txt")

and that should do it too.

[tristan.verhoeff]

Thanks, that seems to have fixed it :)