Measuring cell velocity (or speed) during migration

A quick question regarding quantification of cellular velocity (or speed) of migration on planar surface of a tissue culture flask.

I have a time-lapse set of images taken every minute for a period of 24 hours. The cells are illuminated with ordinary white light and the images are recorded in grey scale and saved as JPEGs. Relative to the off-white background the cells are darker grey. There are typically 10-100 cells in the field of view in the image.

I would like to use CellProfiler to identify each of the individual cells positions in the first recorded image (t=0). I would then like to track each individual cell’s position in the consecutive images which are taken every minute for a total time of 24 hours. Ideally I would like to obtain cellular migration velocity (i.e. speed and direction); however just the cellular speed of migration would also be great.

From this data, I would like to plot a graph with the x-axis being time, and the y-axis being cell speed (or better yet velocity, which would require an x-y-z plot or so). As a result I would be able to visualise the effects of say various drugs and their impact on cellular velocity over time. Also further detailed analysis of the graphs would enable me to compare two different drugs, which say both result in inhibition of cellular migration down to 50% after 24h of incubation; however I would be able to see that drug number one has much quicker mode of action than drug number two. I would also like change the drug solutions back to control media solution at t=24h, and would hopefully be able to see how reversible the cellular velocity is upon removing the drug solutions.

Also, as an added challenge; I sometimes come across lateral drift of the culture flask over a period of 24 hours. So the entire field of view can for example gradually drift right by 100µm over the 24 hours. Obviously this is quite a nuisance and will add error to the measured cellular migration speeds. I have noticed that in one of your example pipelines you can easily deal with object drift.

I would like to know if it is possible to use this function to help me overcome my drifting problems. Sometimes there is a bit of debris around the cells which does not move on the culture dish surface; could I tell CellProfiler to use this as reference point to calculate the drift over time. Or since every cell will drift at the same velocity, could CellProfiler calculate the drift by analysing the cellular migration speeds and estimate the lateral drift of the culture dish. Or perhaps there is another way to detect and subtract the drifting speed of the culture dish that I am not aware of?

Thank you in advance for the effort of helping me with these problems.

Best regards,
Alex

Any sort of tracking is completely dependent on whether your cells can be identified and segmented reliably. If this is so, tracking should be doable with the TrackObjects module. However, since your cells are not fluorescently stained, this could be a problem. Take a look at #3 in the FAQ and see if CellProfiler can handle your cells first:

http://forum.image.sc/t/cellprofiler-2-0-faq/12699/1

While CP can collect the x,y locations of each tracked cell, currently, we do not have a means to plot these results in CP. You would need to export the data as a spreadsheet and analyze it for velocity in another software package (such as Excel or R), or use the output file directly which is openable in MATLAB.

We use the Align module which is able to multiple images up if a reference image is provided which can be used for this application. However, this works if images have a good degree of correlation with each other. Large features (such as the debris you mention) might suffice for this purpose but the motion of the cells hamper the alignment. You can try making a pipeline with the Align module using a couple of images at two time points to see if the spatial features are sufficient for this purpose.

Regards,
-Mark