DAB object classification

I am trying to analyse the intensity of a nuclear protein in cell culture samples using DAB staining. The problem is: the protein appears only after pharmachological treatment so the nuclei of my controls are not stained, but are identifiable by eye (see attachment).
I prepared a pipeline that worked just fine with my test run, but it would not identify the nuclei of the controls in this set of images, so I added a smooth module to enhance the contrast, but it still doesn’t work.

ML_151112_11_pipeline.cpproj (799.5 KB)


Hi! I’m no pro but maybe this helps.

If you have the option of counterstaining with a fluorescent nuclear stain such as DAPI, you could acquire a fluorescent image of the nuclei and detect the nuclei in one channel while measuring DAB in the other. The DAB might quench some of the fluorescent signal in your experimental condition but that would not be a big problem.

If you have the option of acquiring images using phase contrast, you might be able to process the images in a way to detect the nuclei (e.g. invert and blur), but in my humble experience that is not very reliable.

Another option I could think of is using the “IdentifyObjectsManually” module - but tracing each nucleus by hand is a tedius job, to say the least.

All the best!

p.s. As far as I understand CellProfiler is most suited for fluorescent images (bright foreground, dark background) so if you want to make the most of the software (lots of samples, high throughput) it might be worth considering using a fluorescent secondary antibody for your protein (together with a fluorescent nuclear stain). Makes analying nuclear translocations etc a breeze.

Hi, thank you for your suggiestions. @Fabba123 The idea of DAPI seems good, I will try it next time. For the moment, since I have a lot of images to analyse and I don’t want to throw them away, I tricked the process blurring, so that at the end the nuclei are recognized anyway. It’s not elegant, but somehow it does the job. Unfortunately, I need to use two different pipelines for the analysis of the two groups. I know that CP is optimized for immunofluorescence, but my primary antibody refused to work properly in those conditions.

Thanks for the good advice @Fabba123, and let us know if you need further help @Monica_Langiu!