[Bioc-devel] Moderately large files in an Experiment Data package?

[Barry, Timothy P]

Hello all,

I have initiated the submission of three packages to Bioconductor: sceptre<https://katsevich-lab.github.io/sceptre/> (an R package for perturb-seq analysis), ondisc<https://timothy-barry.github.io/ondisc/> (a companion R package to sceptre that implements new data structures for large-scale single-cell data), and sceptredata<https://github.com/Katsevich-Lab/sceptredata> (an experiment data package that provides example data for sceptre and ondisc). ondisc depends on sceptredata, and sceptre in turn depends on both ondisc and sceptredata. Our updated user manual<https://timothy-barry.github.io/sceptre-book/> describes how all three of these packages interface with one another.

In accordance with the Bioconductor submission instructions, I submitted the data package (i.e., sceptredata) first<https://github.com/Bioconductor/Contributions/issues/3386>. However, I received the following error message: "The package contains individual files over 5Mb in size. This is currently not allowed.” Indeed, sceptredata contains two files that are 11MB and one file that is 6MB. The package stores example data in both the `data` directory and the `inst/extdata` directory.

I thought that experiment data packages were allowed to have larger files? If not, does anyone have a recommendation for how I should proceed? Kasper Hansen suggested ExperimentHub as a solution. Might that the way to go?

Thank you greatly for the help!
Tim


	[[alternative HTML version deleted]]