abagen icon indicating copy to clipboard operation
abagen copied to clipboard

Automated re-annotator for AHBA probes

Open rmarkello opened this issue 5 years ago • 0 comments

The issue

As Arnatkevičiūte et al., 2019 pointed out in their manuscript, some of the probes in the AHBA data were incorrectly (or not) annotated to genes. Since the initial data release by the Allen Institute the reference gene database has been updated many times (most recently in July, according to this database), and reannotation with the update database provides new (or different) labels to these probes.

Unfortunately, there's a possibility for changes with every update of the reference gene database, but running the reannotation procedure on your own is sort of a huge pain. As is, we're just using the reannotation released with the Arnatkevičiūte manuscript in abagen, but it might be nice to find an alternative method.

Proposed solution

Create a Docker image (or something) that can be set up as a cron job on TravisCI or CircleCI to run the reannotation procedure every time the gene database is updated (or just once a week or a month or something?) so that our probe-to-gene annotations are constantly "up-to-date."

The biggest pain is cobbling together the different software packages that are required to reannotate the probe data, so having an easy-to-use implementation with Docker would probably be, at a minimum, ideal. But having something that uploads new annotations every so often that we can pull from would be great, too!

rmarkello avatar Sep 20 '19 19:09 rmarkello