bioconda-recipes: bioconductor-chromvar is broken

Hi,

The recently updated ‘bioconda-chromvar’ bioconda package doesn’t work any more. (Its R package works well so I suppose it has something to do with bioconda)

I’m attaching the R script and datasets here to reproduce the errors. https://www.dropbox.com/sh/ts3a6bjrija6qw7/AAD0ryO5qzXvcyO6NKMJQMera?dl=0

$ conda create -n chromvar_test python=3.6 $ conda activate chromvar_test $ conda install bioconductor-chromvar $ conda install bioconductor-bsgenome.hsapiens.ucsc.hg19 $ Rscript chromVAR.R

Several errors will be reported:

Error1 : libssl.so.1.0.0: cannot open shared object file: No such file or directory
It is caused by library mismatches between r-openssl and openssl Fixed by conda install openssl==1.0.2p

Error2: libgsl.so.19: cannot open shared object file: No such file or directory Fixed by conda install gsl=2.2.1

Error3: Error in library(BSgenome.Hsapiens.UCSC.hg19) : there is no package called ‘BSgenome.Hsapiens.UCSC.hg19’ It looks like conda install bioconductor-bsgenome.hsapiens.ucsc.hg19 doesn’t really work Fixed by installing ‘BSgenome.Hsapiens.UCSC.hg19’ within R $ R

if (!requireNamespace("BiocManager", quietly = TRUE))
    install.packages("BiocManager")
BiocManager::install("BSgenome.Hsapiens.UCSC.hg19", version = "3.8")

Error4: Segmentation fault This error happens at the step of ‘kmerdev <- computeDeviations(SE, KmerMatch)’ in R script We tracked down the dependency package 'r-nabor '. Fixed by installing ‘nabor’ within R $ R > install.packages("nabor")

I did my best but it seems beyond my capabilities to fix all the issues. I would be really grateful if you could help fix them. Many thanks!

About this issue

  • Original URL
  • State: closed
  • Created 5 years ago
  • Reactions: 2
  • Comments: 19 (18 by maintainers)

Most upvoted comments

You’re welcome, Luca! All of this would not work at all if there weren’t users who take the effort of reporting bugs and issues, so thanks go back to you guys!

Please refrain from reopening this. The actual issue is being addressed here.

Use the --no-update-dependencies option.

My guess is that that’s a byproduct of you not listing everything you needed in one command. Try conda create -n bioc_debug python=3.6 bioconductor-chromvar bioconductor-bsgenome.hsapiens.ucsc.hg19 to avoid dependencies from getting upgraded incorrectly. I’ve just run exactly that again and once again got an appropriate openssl.