Velocity is calculated on a Frame to Frame basis -> can this be changed?

[howlermonkeys]

Track velocity is calculated frame to frame, which, to the uninformed user, gives a fake report of velocity. (Diffusing particles do not have velocity, or immobile particles will have the noise of assignment as velocity).

Is it possible to incorporate MSD vs t fitting?

[tinevez]

Hi @howlermonkeys Most likely not in the Mastodon directly. You would export tracks in an data analysis software and perform MSD fitting there. Example http://tinevez.github.io/msdanalyzer/

[imagejan]

We've been using TrackMate from within KNIME successfully with these plugins:

https://github.com/fmi-faim/fmi-ij2-plugins/

Having the track data in KNIME makes MSD analysis really easy. For a use case (in this case relative MSD between paired tracks of two channels), see this paper and the attached workflow: http://doi.org/10.13140/RG.2.2.16490.31688

[howlermonkeys]

Thank you. We do our analysis in Matlab, so that is not my concern.

The issue is the "velocity" is being reported by multiple people in papers, (who don't know any better) when this is not velocity. We will try and make an MSD vs t plugin to go on top of trackmate, but it may be a year or so.

But, at least a warning about this velocity, or a public mention of how it is calculated. (We were digging through the lit for a few months).. but found out by contacting the developers.

On Tue, May 9, 2017 at 12:39 PM, Jan Eglinger notifications@github.com wrote:

We've been using TrackMate from within KNIME http://knime.imagej.net/ successfully with these plugins:

https://github.com/fmi-faim/fmi-ij2-plugins/

Having the track data in KNIME makes MSD analysis really easy. For a use case (in this case relative MSD between tracks of two channels), see this paper https://www.researchgate.net/publication/312181809_Histone_degradation_in_response_to_DNA_damage_enhances_chromatin_dynamics_and_recombination_rates and the attached workflow: http://doi.org/10.13140/RG.2.2.16490.31688

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/fiji/TrackMate3/issues/31#issuecomment-300225226, or mute the thread https://github.com/notifications/unsubscribe-auth/ABGFOdgyrmrh8NiznP600RssUBpNvJPUks5r4JbDgaJpZM4NVesL .

-- Ethan Garner - egarner@g.harvard.edu John L. Loeb Associate Professor of the Natural Sciences Harvard - Molecular and Cellular Biology / Center for Systems Biology

Please contact me via digital means rather than analog. To schedule a meeting: http://meetme.so/EthanGarner Free / Busy schedule: https://freebusy.io/howlermonkeys@gmail.com Website http://garnerlab.fas.harvard.edu

[imagejan]

Just to make sure: we're talking about TrackMate and not TrackMate3 (i.e. this repository, aka Mastodon) currently, right?

In TrackMate, I see Velocity being reported for each link between to vertices, and Mean track speed for each track. I see no problem in reporting a velocity for each frame-to-frame link, but of course it should not be confused with a per-track speed. Isn't the reporting clear enough?

[howlermonkeys]

I see where the confusion comes from, and this is often a confused point I come across in paper I review. Velocity is a sustained motion in a directional manner. This is why MSD vs T is used, to separate out the directional vs diffusive components of motion.

I.e. diffusing particles have instantaneous velocity (frame to frame). In some cases the mean of this can result in a true velocity, but only if the particle is moving in a line. But, a diffusing particle is not moving directionally, and thus "mean velocity" gives a misrepresentation, and people get confused between the two.

I think it might be worth reporting as "frame to frame velocity" or "instantaneous velocity", as I have seen many cases where people think that Trackmate is calculating MSDs, or people report this as a true velocity of directional motion., which is not correct. (biologists just assume "I used trackmate, therefore I don't need to know how it works" ).

But overall, it is unclear, and we had to email the developers to find out how it worked... because before this we were digging through the lit... (and arguing amongst ourselves and collaborators) ... and my students were running multiple tests on directionally moving particles to prove to me something was not accurate.

On Wed, May 10, 2017 at 8:51 AM, Jan Eglinger notifications@github.com wrote:

Just to make sure: we're talking about TrackMate https://github.com/fiji/TrackMate and not TrackMate3 (i.e. this repository, aka Mastodon) currently, right?

In TrackMate, I see Velocity being reported for each link between to vertices, and Mean track speed for each track. I see no problem in reporting a velocity for each frame-to-frame link, but of course it should not be confused with a per-track speed. Isn't the reporting clear enough?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/fiji/TrackMate3/issues/31#issuecomment-300472107, or mute the thread https://github.com/notifications/unsubscribe-auth/ABGFORNW74JMjUNAULnCF1bJFOEeZEKmks5r4bLcgaJpZM4NVesL .

-- Ethan Garner - egarner@g.harvard.edu John L. Loeb Associate Professor of the Natural Sciences Harvard - Molecular and Cellular Biology / Center for Systems Biology

Please contact me via digital means rather than analog. To schedule a meeting: http://meetme.so/EthanGarner Free / Busy schedule: https://freebusy.io/howlermonkeys@gmail.com Website http://garnerlab.fas.harvard.edu