A Common Format for Worm Tracking Data
Michael Currie (1a), Rex A. Kerr (2), Chee Wai Lee (1b), Jim Hokanson(1c), and Andrew E.X. Brown (3).
- OpenWorm (a. Somewhere; b. Somewhere Else; c. Somewhere Yet Again)
- Calico Life Sciences
- Imperial College London
Many labs, including ours, have built a wide variety of worm trackers. These have a wide range of capabilities, from high-resolution imaging of single animals during calcium imaging, to very low-resolution imaging of animals as points. This diversity of capability enables the C. elegans community to address a wide range of problems at an appropriate scale.
Most of these trackers also produce some data that is very similar to that of other trackers: animal position or spine, for example. Unfortunately, each tracker uses its own format to store data, so that any later analysis, despite being general in nature, cannot be performed on data from different machines. As the volume of tracking data grows, and the variety of downstream analysis methods expands, this limitation will pose an increasingly large barrier to replication of and extension of existing work across different labs.
To address this issue, we have defined the Worm Common Object Notation, a set of rules for how to write tracking data in the ubiquitous JSON format, so that it can be easily shared between labs. To facilitate easy adoption of WCON, we have further written software in a variety of languages that will read or write data in WCON format. So far, we have implementations in Python, Scala, Matlab, and Julia, and wrapper libraries for Octave, R, and Java to use one of the main implementations.
Additionally, the Tracker Commons project of which WCON is a part contains a small but rapidly growing set of pre-packaged analysis tools for routine manipulation of worm tracking data. We will also maintain a list of other WCON-compatible analysis tools as they become available.
If you are involved in worm tracking, we invite you to adopt WCON and help make C. elegans behavioral data widely accessible. WCON is developed under the open source Tracker Commons project of the OpenWorm Foundation. We invite contributions and improvements!