trackObjects TrackObjects_Label and TrackObjects_TrajectoryX

Hello,

I tried to trackobjects with time lapse picture.
I had some questions about this pipeline:

  1. The trackObjects_label observed on the picture corresponds to Nan trackObjects_trajectoryX and Y (for some of them)
  2. trackObject considered mitosis whereas cells merged and another number of cell was assign instead of the same number
    Any suggestions will be appreciated

Hi,

  1. This typically happens if you set a minimum persistence length for a track (if you say you only want tracks that last >5 frames, for example, the Label from tracks that only last 1-4 frames will be set to NaN)- can you verify that that’s the case in your pipeline?

  2. You can use a tracking method other than LAP if you don’t want the mitosis score to be calculated- in general, calibration of LAP can take a while to get right (but is helpful when your objects are dim and/or crowded). Is this a problem that’s happening frequently, or just in one particular case? In general it’s extremely difficult to calibrate any automated tracking algorithm to 100% accuracy.

Thank you for your answer.

  1. I set “maximum temporal gap, in frame” = 6.
    Some cells merged and disappeared for 4 to 6 frames and then the second cell reappeared after 4 to 6 frames and another number was assigned whereas the same number of this cell. It look like a mitosis and not a merge of cells (see cells 5 and 7 for example) . How can I set merging cells in this case (in attached trackObjects)?
  2. Some cells divided, thus I set a “mitosis alternative cost” to 360.

Hi,

If you’re using the second phase of the LAP algorithm, the pictures aren’t an accurate depiction of the final tracks, as they’re output only after the FIRST phase- the second phase is run later. This is mentioned in the module help.

There’s some very good advice in the TrackObjects help about how to optimize all the settings, I recommend taking a look at that to help you further optimize!

Good luck :slight_smile: