How to encode root/parent relationships in pollen growth data

[cedarwarman]

Hi, I'm not sure if this question is a little too distant from the typical use case of cell-tracking, but I figured I would ask.

I am tracking pollen grains and pollen tube tips in images like this:

Right now I'm tracking tube tips and pollen grains separately, but I would like to associate each tube tip with a pollen grain (as the parent, if I'm using the terminology correctly). I'm wondering if there's a way to encode this behavior in BayesianTracker. Pollen behaves a little different than other cells, it would have to follow these rules:

• Each tube tip track needs to have 1 pollen grain parent
• Pollen grains can have up 3, but usually 1, tube growing out of them (child nodes)
• Pollen grains and tube tips are distinguished in the data by their class, which I've added to the PyTrackObject as an extra property. Pollen grains can be one of several classes (e.g. ungerminated, germinated, burst) while pollen tube tips are always the pollen_tube_tip class.

Thanks for your help!

[quantumjot]

Sorry I didn't get back to this - I think this particular question might be challenging to address with btrack because of the way it works. However, we have another project that might be interesting for this: https://github.com/alan-turing-institute/grace