Workshop Session - Image Quantification Workshop

Preliminary Set-up of Fiji

Note: You must have an open image to do this.

If you don't have FIJI on your computer - search for "FIJI imageJ". Download/install.

• Look at full 3-color image to see context of what we are working with.
  • Open image: BPAEfull_R3D copy.dv by dragging it onto the Fiji icon, or using the File ----> Open menu. If this doesn't work or look right - try this --> Plugins/LOCI/BioFormats Importer under color options pick "composite". --> look at image --> close
• Always work with a copy of your image - never the original!
• Open image: BPAE32nd_R3D_w435 copy, Image ---> Adjust ---> Brightness/Contrast... in the box that appears, click on the slider above the word “Maximum”. Move slider to right. This will have an effect on some images in some version of the software. This image is just the blue channel of the full image.
• Analyze ---> Set Scale... Click to Remove Scale, select Global, then click OK. This removes previous scale settings that may be wrong.
• Analyze ---> Set Measurements... choose, Area, Standard deviation, Center of Mass, Integrated density, Mean gray value, Display Label, Decimal Places =1, then click OK. This picks what measurements we record.

What we want to do is measure the level of chromatin signal in the 5 cells and compare them to each other. DNA was stained with Hoechst stain.

• Set-up regions of interest (ROI's). click on the nuclei image to select it Process ---> Filters ---> Gaussian Blur Sigma (Radius) = 10, click OK. We want to pick the whole nucleus - so blurring smooths small differences.
• Image ---> Adjust ----> Threshold.... select the Otsu method, Dark Background, click Auto, but don’t click Apply. Nuclei should all be smoothly covered with uniform red color
• From the Fiji tool bar, select the “Wand (tracing)” tool, and click on the topmost red nucleus. A yellow border should appear around it
• Edit ---> Selection ---> Add to Manager a new small window labeled “ROI Manager” should appear, with one item showing on the left, something like “0001-0158-0337”
• With the wand tool, click on the next red nuclear mask, and in the ROI Manager box click “Add [t]”. Repeat for all the remaining nuclei.
• When all have been entered, select all of the ROI’s in the list, click More>> Save... on the ROI Manager box, and save the list of ROI’s to a disk file. It will be used for measurements on both the uncorrected and corrected images
• Click off the wand tool - it can cause trouble later.
• Close Image --> DON'T SAVE

• Reopen image: BPAE32nd_R3D_w435 copy (you should have it downloaded from above). Adjust the Brightness/Contrast as above.
• In the ROI Manager box, select all of the ROI’s, then click :Measure” This should create a measurement window with the values you selected at the beginning of this exercise. Look at the “IntDens” values for each nucleus. How do the values compare to each other? How much difference is there between the lowest and highest values?

• Open the “Darkcurrent” image (dark_R3D_w435 copy). Adjust the Brightness/Contrast as above. This is an image taken with no light to the camera and is the background noise. Move your pointer around in this image. The intensity of the pixel the pointer is over is shown as value = xxx in the FIJI tool window. Note this value. You may also take another measurement of your ROI's on this image.
• Smooth this image with a Gaussian filter of radius 3 Process ---> Filters ---> Gaussian Blur Sigma (Radius) = 3.0, click OK. We blur to get rid of individual pixel differences. These aren't meaningful - our resolution isn't that good.
• Open the “FlatField” image (flatblue_R3D_w435 copy), adjust Brightness/Contrast and smooth it in the same way. This is an image of a uniform field of fluorescence. If the camera was perfect - the image would be uniform - is it?
• Subtract the smoothed Darkcurrent image from the smoothed Flatfield image
  • Process ---> Image Calculator
  • Image 1 = smoothed Flatfield, Operation = subtract
  • Image 2 = smoothed Darkcurrent, Operation = Subtract
  • Check Create New Window and 32 bit result --> then click OK
  • Save the Result as a tiff file File ---> Save As... ---> Tiff...
  • Name it something memorable, such as FminusD.tif
  • Get the average value of this image and write down the number.
  • Analyze ---> Histogram , Write down value for “Mean”.
  • Measure ROI's for this image. How much variation do you see?
• Close the Flatfield image; DON’T save the changes.
• Open the nuclei image (already downloaded from above). Adjust Brightness/Contrast Subtract the smoothed Dark current image from the nuclei image. Rename, measure
• Divide the Result by the FminusD image Process ---> Image Calculator i1 = Result, i2 = FminusD, Operation = Divide, mutiply by mean you wrote down above. Process --> Math --> Multiply Then fix brightness/contrast. Image --> Adjust --> Brightness & Contrast --> click Auto Save the Result as a tiff file
• In the ROI Manager box, select all of the ROI’s, click on the “Result” image and then click :Measure”

In the “Results” window, compare the “IntDens” values for the nuclei before and after correction.

Copy measurements to Excel, graph data for 5 nuclei for each of the image treatments. What can you say about the amount of DNA in each nucleus? How do the "before" and "after" numbers compare? Does the "answer" change? What is your biological hypothesis to explain the differences seen in the "after" data?

If you are feeling confident - try this again with the BPAEfull_R3D copy.dv image (when opening - set to colormode = greyscale, split channels. You can close red and green channel images. Save blue channel as a tiff. What differences do you see before and after correction? Why?