I'm using the following CDF file for the Affymetrix chip of Brachypodium dystachion. Apparently, it follows all the rules of the Affymetrix CDF specification
http://giorgilab.org/BradiGiorgiMutwil001_txt.cdf
However, when I try to create a customcdf environment out of it, using R and the Bioconductor package makecdfenv
library(makecdfenv) make.cdf.package("BradiGiorgiMutwil001_txt.cdf",species="Brachypodium_dystachion")
I get the following error:
Reading CDF file.
* caught segfault * address 0x1, cause 'memory not mapped'
Traceback: 1: .Call("readCDFfile", as.character(file), as.integer(3), as.integer(compress), PACKAGE = "makecdfenv") 2: read.cdffile(file.path(path.expand(cdf.path), filename), compress = compress) 3: make.cdf.env(filename, cdf.path = cdf.path, compress = compress, return.env.only = FALSE) 4: make.cdf.package("BradiGiorgiMutwil001_txt.cdf", species = "Brachypodium_dystachion")
Is there anything that strikes you as wrong in the input CDF file? I'm totally puzzled, because I don't know how to interprete this segmentation fault, nor how to trace it to a specific problem. Something similar happens while using the affxparser Bioconductor package, so it must be a problem with the CDF fields (and not with the package).
Thanks a lot! :-)
Federico
The function make.cdf.package() does not require a binary cdf, so there is no reason to have a test.
I took a look at this yesterday, and there isn't anything obviously wrong with this cdf, so it's nice that simply converting to binary fixes things.