Problem with consistency in 53BP1 foci detection

53bp1_count.cppipe (10.4 KB)

Background

I have acquired images with 3-channels: DAPI, EdU (S phase marker), 53BP1. In regards to the microscope used, these images were taken on an advanced spinning disk confocal microscope, 40x/1.3NA objective, with 16-bit depth camera and 1x1 binning. I have acquired a Z-stack (~11 planes, a total of 5um), for which I have done a maximum intensity Z projection so it is easier for analysis.

Analysis goals

I wish to identify the number of 53BP1 foci for each cell. However, the number and size of these foci differ between cell cycle stages. For this reason I also wish to measure the intensity of DAPI and EdU channel for each cell, to create a cell cycle profile, from which I can separate the 53bp1 foci between G1/S/G2 stages.

Challenges

I have created a pipeline using online guides which can identify the 53BP1 foci, however I found that it is very inconsistent between individual images. In one image, the foci identification would be close to perfect, how I would count them if I were to do it manually (see two good examples I have attached). However, when I tried to check another image set, the foci identification was less good (I have included two bad examples). I tried to change some of the thresholding parameters, but they would apply just to one image set, then they would be useless when I would move to another image set. I cannot figure why this is the case, as the images were all acquired using the same microscope settings.
I have attached the pipeline and two original sets of images: Set 1 worked nicely in identifying the foci, whilst set 2 did not.

Extra information:
The original images were in Slidebook files (.sld), which I have used to make the Max intensity Z projection and then to export as .tif files.

If anyone has any clues it would be great help! :raised_hands:

I have also attached the two original image sets below. [1/2]

Set1_C0.tif (8.0 MB) Set1_C1.tif (8.0 MB) Set1_C2.tif (8.0 MB)

I have also attached the two original image sets below. [2/2]

Set2_C0.tif (8.0 MB) Set2_C1.tif (8.0 MB) Set2_C2.tif (8.0 MB)

Hi @CameliaM,

I checked you pipeline & images. Your pipeline is perfectly is fine expect for few parameter setting. Once that is changed, it works fine. The change I made is,

  1. Changed the size of the speckles in “EnhanceSuppressfeatures module”. It should the similar size of your speckle.
  2. I changed your minimum threshold value instead of Zero.
    It works fine for both the images. There some few speckles it is picking up but it is not considered as object because of size. In case you want to include them you might have to change the size cutoff.
    PFA screenshot & pipeline.

    53bp1_count_lb.cpproj (658.0 KB)

Hope this helps.

Regards,
Lakshmi

Hi @lakshmi,

I have tried your pipeline and indeed it works very nice!
However, I noticed in my analysis some weird results in some of the images, and I went back to run them individually. It seems that again the threshold or something else is not fitting well with those image sets. I have included one example which gave me big problems. I am not sure whether it is the Enhance module which does this. I was considering dropping it altogether and just masking directly on the image. What do you think?
Thanks a lot for the help.

CameliaSet3_C0.tif (8.0 MB) Set3_C1.tif (8.0 MB) Set3_C2.tif (8.0 MB)

Hi @CameliaM,

Ok. Good.

I am not sure exactly what is the problem. I checked with the set3 image, in few nuclei I found that the spots are not identified completely & missed from the Enhance module. Just while removing the background some signal also being removed (in few spots). I tried changing parameters in Enhance module & it improved a bit. Most of them are correct except for very few. Please check the attached pipeline.

53bp1_count_lb.cpproj (658.0 KB)

Regards,
Lakshmi
www.wakoautomation.com