Normalisation for fluorescence intensity from images acquired at different time points

cellprofiler
normalization
image-cytometry
time-series

#1

Hello,

My experiment involves acquiring images at different time points and the analyse the fluorescence intensity from single cell. Note that the image frames taken at different time points are not the same. The expression of the protein is time-dependnet. I would like to plot a quantitative dot plot of time-dependent protein expression in single cell. In this case how do I normalise for fluorescence intensity from images taken at different time points? Is there any image analysis pipeline in CellProfiler? Any help would be much appreciated.

With regards,
Vimal


#2

Hello Vimal,
I might be misunderstanding your question but it seems like you should be able to look at intensity at different time points in the exported spreadsheet. If you have a set of images representing different time points you should be able to obtain intensity in locations of the segmented objects in each image that is represented by a different number in the exported spreadsheet. Then you can make a dot plot by using that intensity data.
Best,
VC


#3

Hi Cherny,
Thank you. Yeah, I have the time-dependent intensity of objects. But those images were acquired at different time points, which means I need to normalise them before plotting them together. Since the illumination beam intensity and other factors controlling the optics of the microscope vary on a day-to-day basis. If I do not normalise them, the data would completely mislead my experiment and I would be interpreting erroneous data. So in this case, what would be the ideal way to represent my data?


#4

Hello Vimal,
I would use the RescaleIntensity module at the beginning of the pipeline and try out the available rescaling methods to adjust for the variation in intensity from image to image. If you are still having trouble and don’t mind sending some of your images with draft pipeline I can try optimizing it.
Best,
VC


#5

Hi Vimal,

If you want to correct for hardware variations day to day, you need to have some sort of a reference standard as part of your experiment, such as a set of beads or something else that wouldn’t be expected to bleach much, or to have measured those variables as you went. Otherwise the best you can hope for is a relative normalization to another control within the experiment- ie on day 1 when I took pictures of all of my cells well 2 was twice as bright as well 1, but on day 2 well 2 was three times as bright as well 1, etc.


#6

Thank you so much! I would definitely try the first one!