Identify mitosis

Hi,

I’m a new to cell profiler and cell profiler analyst.
I’m doing a project at university.
I have 208 pics of cells over 52 hours, during that time the cells are multiplying-doing mitosis. I want to identify and to be able to count how many mitosis happened through that 52 hours.I there any way i can do it?

Thank you,
Hilla

(Moving topic to appropriate board…)

Hi Hilla,

Your description is somewhat vague, unfortunately. Without additional information on the assay, we cannot tell whether CellProfiler will be of help to you. Please read our pre-posting FAQ, provide the details described, and we can proceed from there.

Regards,
-Mark

hi again, i will try and make myself clearer.
I’m attaching 4 photos (in total I have 209 photos), this photos were made in a 15 minutes break from one another during that time some cells are splitting or merging.
I’m also attaching my pipeline.
My goal is to make a list of trajectories-every cell that has been splitted I want to know what was the first frame the cell appeared in, on what frame did the cell splitted, how many frames were from when the cell appeared till it splitted, i want this to be a trajectoriy. I want to count how many of them I have.
Is there any way I can do it on cellprofiler?

I hope I was clearer,
Thank you,
Hilla
DefaultOUT__145.mat (986 KB)








Hi Hilla,

You pipeline appears to basically collect the information that you want. However, the the various cost settings in the LAP method in TrackObjects requires a decent degree of tweaking in order to make sure that splits are caught correctly. Unfortunately, these computations are perform after the last cycle of the pipeline, and hence aren’t available for display to the user as a sanity check.

The only way we have found to do this is to use a third-software package (like MATLAB) to read the x-y locations from output csv spreadsheet and plot them for visual review. Clearly this approach is sub-optimal for what you want to do, but it is currently the only option.

However, assuming that the tracking is performed correctly. the metrics you want are indeed in the output csv. But again, you would need third-aprty software to pull the requisite info from the file. I can describe the basic workflow if you were to go this route, but actually devising the code to this in whatever software you choose is up to you:

  • First frame of cell appearance:
    Each cell is given a label at the outset, if you look in the per-object file under the header containing"TrackObjects_Label". Find all unique labels, then find the ImageNumber (i.e., frame number) in the row that corresponds to the 1st appearance of that label. - Frame in which a cell split:
    As it stands right now, if an object splits, the progeny inherit the label of the parent. So for each unique label, you could search the file for rows in which multiple occurrences of the same label occur in the same frame; this would indicate that a split has occurred. If you find a row where this occurs, find the ImageNumber associate with that row. - Number of frames from appearance to split:
    Once you have the above two values, subtract them to get this duration.
    Regards,
    -Mark

Hi Mark,
thanks for your quick response.
Do you already have a code for this in MATLAB? it would really help me.

Thanks,
Hilla.

Hi Hilla,

I’m posting some MATLAB code; it reads in a CellProfiler-generated .mat file, tries to sort the enclosed object names and make 3-D plots. But please note (and I cannot emphasize this enough) it’s very basic and completely unsupported by me or anyone else.

Regards,
-Mark
plot_time_series.m (4.17 KB)