Please use files in https://github.com/comorment/containers/src instead.
This repository is used to develop and document singularity containers with various software and analytical tools for GWAS and post-GWAS analysis.
For new users we recommend to go over introductory instructions in docs/hello.md, which explain the basic usage of singularity containers, using a minimalistic example (singularity container with plink
binary).
If you would like to contribute to developing these containers, please see docs/contributing.md.
For a tutorial on GWAS with synthetic data, see docs/gwas.md.
- download containers shared on the Google Drive.
- download
comorment_ref.tar.gz
file from the above Google Drive folder, extract it withtar -xzvf comorment_ref.tar.gz
command, and create an environmental variableCOMORMENT_REF
pointing to the folder containing extractedcomorment_ref.tar.gz
data. If you want to see the content ofcomorment_ref.tar.gz
without downloading and extracting, you may take a quick look here. This is a private repository, and you need to get access. Please contact Oleksandr and Bayram by e-mail and send us your github user name. If you don't have it, create one here. - create an empty folder called
data
, for storing the results and intermediate files produced by running containers. (most instructinos mount this folder like this:-B data:/data
).
hello
- a hello-world introductory containergwas
- basic tools for gwas (plink
,plink2
,prsice
,BoltLMM
)python3
- python3 packages distributed via miniconda. This package also contains jupyter notebook.ldsc
- LD score regressionmatlab
- container allowing to run pre-compiled MATLAB software. This container also has OCTAVE installed.R
- contaienr for R analysis (installed by native R package manager)
All containers have a shared layer of common utilities (wget
, gzip
, etc).
If you face any issues, or if you need additional software, please let us know by creating an issue.
We use NREC machine to develop and build containers.
NREC machine has small local disk (~20 TB) and a larger external volume attached (~400 TB)
If you use NREC machine, it's important to not store large data or install large software to your home folder which is located on a small disk,
using /nrec/projects space
instead:
Filesystem Size Used Avail Use% Mounted on
/dev/sda1 20G 9.6G 9.7G 50% /
/dev/mapper/nrec_extvol-comorment 393G 346G 28G 93% /nrec/projects
/dev/mapper/nrec_extvol_2-comorment_2 935G 609G 279G 69% /nrec/space
Both docker and singularity were configured to avoid placing cached files into local file system.
For docker this involves changing /etc/docker/daemon.json
file by adding this:
{
"data-root": "/nrec/projects/docker_root"
}
(as described https://tienbm90.medium.com/how-to-change-docker-root-data-directory-89a39be1a70b ; you may use docker info
command to check the data-root)
For singularity, the configuration is described here https://sylabs.io/guides/3.6/user-guide/build_env.html and it was done for the root user by adding the following line into /etc/environment
export SINGULARITY_CACHEDIR="/nrec/projects/singularity_cache"
Common software, such as git-lfs, is installed to /nrec/projects/bin.
Therefore it's reasonable for all users of the NREC comorment instance
to add this folder to the path by changing ~/.bashrc
and ~/.bash_profile
.
export PATH="/nrec/projects/bin:$PATH"
A cloned version of comorment repositories is available here:
/nrec/projects/github/comorment/containers
/nrec/projects/github/comorment/reference
Feel free to change these folders and use git pull / git push. TBD: currently the folder is cloned as 'ofrei' user - I'm not sure if it will actually work to pull & push. But let's figure this out.