Docker/singularity container?

[zktuong]

How does one set this up? hmm.

[grst]

I'd recommend you to create a bioconda recipe. Once you have that you get a container for free through biocontainers.

[zktuong]

Singularity container for v0.1.0 all set up and ready for testing!
https://cloud.sylabs.io/library/kt16/default/sc-dandelion

singularity pull library://kt16/default/sc-dandelion:latest
singularity shell sc-dandelion.sif

Just earmarking some not-so-distant plans:

• streamline preprocessing

• @ktpolanski proposed to streamline the pre-processing step to something like a module/wrapper so that a user can essentially do something like:
  container run dandelion-preprocess -o [OPTIONS] on/this/folder/here

• This should solve some of the installation issues that required 'advanced' knowledge of how to set up a working R, python environment that pointed to the dependencies and reference databases.

• Reference databases and environmental variables are also set-up so no need to mess around with profile scripts.

• Can add in a option to specify user-supplied database to overwrite the container's

• This needs a clever system to do multi-sample preprocessing, especially for the reassign_alleles step

• current proposal is to have a text file containing columns specifying the sample id, prefixes/suffixes to add to the fasta and a column indicating if a sample is meant to be combined for reassign_alleles or not.

• trim back on softwares

• the image might be too big (1.5gb) and i should look into removing some potentially unnecessary softwares for the build.

• I think it probably won't change much though because gcc is required to build the R packages.