Posted on

Counting cell colonies with ImageJ and analyzing the data with graphpad

We standardised a way of counting cell colonies from CFU experiments with the ImageJ software. It is important to have the colonies well stained with crystal violet and photograph the plates with a camera from a fixed height.
Photos should be then, treated as follows:
  1. Adjust the color threshold of the image on Image > Adjust > Threshold
    1. Hue: 163-255
    2. Saturation: 46-255
    3. Brightness: 0-255
    4. Thresholding method: Default
    5. Threshold color: B&W
    6. Color space: HSB
    7. Then convert the image to binary on Process > Binary > Make binary
    8. Fill holes on Process > Binary > Fill holes
    9. Separate colonies on Process > Binary > Watershed
    10. Save it as binary
  2. Count colonies on Analyze > Particles; Parameters here may vary with the cell line
    1. For G361 colonies we usually count colonies with a size (pixel2) of 100 – Infinite; a cicularity of 0 – 1
    2. For A375 colonies we usually count colonies with a size (pixel2) of 200 – Infinite; a cicularity of 0 – 1
    3. To optimize the counts one can:
      1. Identify a few colonies with the minimum number of 50 cells and adjust the size value
      2. Count manually the colonies using the multi-point selection tool and then analyse particles until get similar colony numbers
    4. Save the Summary as a txt file and open it as a spreadsheet to get the results
  3. Transform the counts dividing them by the number of seeded cells to get the plating efficiency
  4. Transform the plating efficiency dividing all values by the average of the controls’ triplicates to get the relative plating efficency
  5. Perform an ANOVA with a Bofferoni multiple comparison as a post-test on the relative plating efficiency to find out if differences between treatments are significant
    1. Paste data on a “Column” type spreadsheet, each column being a treatment/experimental condition and each line being a replicate
    2. Perform the ANOVA clicking on “Analyze” icon
    3. Choose on “Column analyses” “One-way ANOVA”
    4. Choose then “Repeated measures ANOVA” as a test and “Bofferoni: compare all pairs of columns” as a post-test, select the column of descriptive statistics for means and averages
    5. If there are significant differences, reflect them on the graph.