Threshold and counting particles


I am working with Stress granules in plants. The componentes of the SGs are usually solubles, however after a heat shock, these componentes start to accumulate forming granules. To quantify my samples, I want to count the number the stress granules induced after a heat shock, however, not always is easy because some of the fluorescence remains, increasing the background and making more difficult establishing the proper threshold. I would like to count the number of granules induced and mesuare their area…

Could you help me? Thank youMAX_HS TROV Plant 1 Stack 1.czi - C=0.tif|attachment (4.0 MB)

Hi Casla,

i think getting counts is relatively easy. I would start out with a Gaussian Filter + Rolling Ball Background Subtraction combined with the Maxima Finder. Here is a description of what to do in ImageJ:

But for the area measurement I would be really careful. Your objects are close to the resolution limit. With a lot of spots you might just measure the size of your PSF.



Thank you for your answer!

What is PSF?

Its the Point spread function of your microscope. A point source of light smaller than the resolution limit will be resolved by the optics of your microscope as an airy disk in xy and like an hourglass in xz. The size of the PSF depends on the resolving power of your system.

So lets say your systems PSF is 250 nm in size in xy. An object larger than that will be fine, although for accurate measurement, it should be a bit larger than that.

However objects that are 100 nm or even 10 nm will still appear as 250 nm in size. The object could be infinitely small, given infinite brightness it would still appear as an object of 250 nm.

So if your objects are close or around the resolution limit you would systematically overestimate the size. If your treatment significantly enlarge your objects this might not be an issue. But a significant reduction in size you will not be able to pick up.

1 Like