Where/how should processed data be deposited when raw data was already published?

Question

[TFTF]

As part of an accepted paper I would like to share both the code and data files so readers can reproduce my results. The code is easy to share via a public repository (e.g. GitHub). However, I am unsure about how to share the data. This is a computational project, so we use published genomic sequence data, process it (mapping/filtering etc), and analyze. I would like to provide the users with the processed data files prior to the downstream analysis (the downstream analysis can be run using the code I will provide). So my question is:

Where can/should I deposit the processed data files for sharing?

Many public scientific data repositories declare that they only accept new data which has not been published already. But the datasets I use have already been published (at least as raw data, or processed with a different pipeline).

Just to be clear, the motivation here is to save users the need to reprocess the entire raw data, which could require significant effort and computational resources.

Answer 1

[Devon Ryan]

We post the datasets we use for Galaxy trainings (fastq, bam, bigwig, vcf, etc.) to Zenodo. This is generally convenient enough, since it's free.

[TFTF]

Thanks! It does indeed seem to be allowed in Zenodo, and there are also nice features linking to GitHub.

Answer 2

[GenoMax]

You can use CyVerse.

[rlwalls2008]

See http://datacommons.cyverse.org/.