I am a beginner with ImageJ and would like to know how to quantify fluorescence images of platelets taken with a fluorescence microscope at 40x and 100x magnification. If you could provide detailed steps for the quantification process and explain what to pay attention to when exporting image files in Thunder, I would greatly appreciate it. Thank you.
The general workflow for counting fluorescent objects in ImageJ is:
Steps 2 and 3 have a large number of options that you will need to adjust to get the best segmentation of your objects of interest.
If you can point out which objects in the image you want to count, I can give you a better starting point.
John Hardy Lockhart thanks, I would like to refer to the following paper on platelets: The endogenous antimicrobial cathelicidin LL37 induces platelet activation and augments thrombus formation. I hope to calculate the fluorescence expression of platelets in the images.
Here's a workflow using the images you provided in the question. I would recommend exporting the images from the microscope without a scale bar and with pixel scaled in microns to make this a bit easier.
1. Correct image pixel scale (if needed)
a. Draw a horizontal line along the scale bar (hold shift while drawing to lock it to 45˚ increments).
b. Analyze > Set Scale... to change units (I got 6.53333 pixels/micron)
2. Separate image into individual channels (if needed) (Image > Color > Split Channels)
a. Close (red) and (blue) images.
b. Duplicate green channel image to use for thresholding/segmentation.
3. Threshold fluorescent pixels from background (Image > Adjust > Threshold...)
a. I used 32 -254 (to exclude scale bar annotation, though there is a shadow of it that I can't avoid), Otsu method, Dark background.
4. Segment masked objects (Analyze > Analyze particles...)
a. I used 5-Infinity micron^2, 0 - 1 circularity, Add to Manager
5. Measure the fluorescence intensity of each object.
a. Select the unused green channel image.
b. Select all objects in the let panel of the ROI manager.
c. Click the 'Measure' button on the right side of the ROI manager.
d. You can change the reported measurements for each object by changing the options in Analyze > Set measurements...
Using these steps, I found 1082 objects in the 80_DM2_ch00.tif image. I think some optimization in the thresholding and segmentation steps could remove some of the smaller objects that I'm not sure are 'real'.
John Hardy Lockhart Thank you for your reply, it helps me understand the concepts of operating ImageJ. I am very grateful.
Thank you for asking this question Lucille Huang, and thank you for the detailed answer John Hardy Lockhart . To piggyback on your question, I have a not-so-similar case where I have a film deposited on the bottom of a beaker and I am trying to determine the surface area of my film but I need imagej to be able to accurately identify what is a film and what is not and I have been using threshold features but I am not confident that it is accurately differentiating my background from the foreground. I am also new to Imagej, and I do not know if there are other features I can use on Imagej to better process my images. Thank you as I await your response John Hardy Lockhart
Damilola Petinrin, it's hard to give any advice without seeing the images you're working with. There is an option to do color thresholding in imageJ, but that is a little more work than the grayscale thresholding done on these fluorescent images. I would recommend you submit it as a new question here on ResearchGate to make it easier for others to find.
- Badges
- Science topic
- Similar topics
- Filing