Help in identifying vessels

Hello I am not sure if anyone can help me or at least guide me in the right direction. My current dilemma is trying to figure out how to quantify the amount of a protein in a pig brain blood vessel. My goal is to find if this protein is increased in diseased pig brain vessels or if it is in fact just an increase in vessels themselves compared to control and drug treated. My protein that I stain for in the vessel increases in intensity with increasing protein amounts. I want to first quantify the amount of stained protein found in the vessels in each picture. I then want to take those same pictures and quantify the amount of vessels regardless of staining intensities. The vessels tend to range in size from small circles (because they are coming at me) and lines. I am having trouble narrowing down the exact modules to use to find such a quantification. I have tried using a few and I can not seem to have the cell profiler differentiate between the hematoxylin stained nuclei,which come out purple and the brown stained vessels containing my protein stained with DAB, which is what I want quantified. any suggestions will help I am definitely not computer savvy so I am having some trouble. I have attached a color image that may help some answer my question.


I’m attaching a pipeline which may help you get started. The key here is using UnmixColors which has DAB as a preset option to obtain an image that can be used as input for feature detection. I also process the unmixed image by performing an illumination correction to remove spatial heterogeneities created by the imaging optics. However, I’ve heard that some caution must be used when quantifying intensities from corrected/normalized images so you may want to check the histological processing literature for more details. Afterward, the intensity from the detected vessels are measured and output as a per-image measurement; if you want per-vessel measurements, you’ll need to add additional modules.

Please keep in mind that this pipeline is to get you started. Some settings may need to be tweaked/optimized for the remainder of your images. You can always refer to the help to the right of each setting for more details.

2010_09_29.cp (6.18 KB)

Thank you very much this has gotten me much closer then before. It is much appreciated!

I am still having much trouble seperating the Nuclei stained with cresyl violet from the vessels stained with DAB. Any suggestions on that?

Hmmm, try the attached pipeline instead. Rather than using color unmixing, I’ve tried color division to accentuate the vessels. See the module notes at the top of each module for details.
2010_10_06.cp (6.35 KB)