Sponsored by the AIRR Community

How to store and share protocols?

Hello everyone–

How shall we host protocols for generation and analysis of repertoire data? I’d like to see if we could collect together thoughts and come to a conclusion. See this thread for some thoughts.

Here are the options I can see so far:

A digital repository

For example, https://figshare.com/ or http://datadryad.org/

  • + public and durable
  • + offer DOIs
  • - expensive and overkill for hosting text documents, if not expensive then the sustainability plan isn’t clear
  • - dispersed among all the material that appears at that site; would need a separate “index” that would keep track of those files
  • - would require effort from investigators, and another account
  • - would be annoying to move to another service should one shut down

This forum

Just upload by making a post.

  • + existing solution integrated into community consciousness
  • + no additional cost
  • - not really designed for the task
  • - requires additional effort from investigators
  • - links can go out of date (though we aren’t going to be moving from our current storage solution anytime soon)


To be clear, this is what I’m imagining. Labs volunteer to have their protocols made public in PDF form. Every four months, a curator blasts out an email to those labs asking for updates to their lab protocols. They reply with PDFs if there are updates. These are then updated to a central GitHub repository in PDF form as well as a plain text version converted from the PDF. The plain text version is mostly there just to have a file that GitHub can “diff” so that users can see differences between versions.

  • + free and centralized
  • + differences between versions are evident
  • + can be very easily moved to another service in the unlikely event that GitHub disappears
  • + minimal effort required on the part of labs
  • - some effort required from a curator
  • - no per-item DOIs (although the central repository could have one)

Thoughts? Would your lab volunteer to have your protocols public?

I would suggest the last option and strongly recommend using Markdown. It is almost impossible to track changes in PDF files + markdown is far easier and intuitive than TeX. We have some of our protocols in markdown format here: https://github.com/repseqio/protocols

Mike, I totally agree that in an ideal world we would be able to get participating labs to put their protocols in a more useful format. In fact, if it was up to me I’d suggest we move to protocols.io. However, in discussion with people running labs this isn’t going to happen.

My thought was that everyone can export PDFs and so we could get a broad diversity of protocols. The text file conversion is just to give some suggestion of diffs between versions. You could totally submit yours in Markdown, which would get extra-credit stars for a single useful format.