I have 2 treatment conditions. I am trying to show that in one condition my protein is randomly distributed in the cytoplasm and in the other condition is at the membrane. We are taking individual z-sections where we can find the edge of a cell not near others. The same rectangle is drawn at 90degrees across the periphery with the midpoint at the membrane. We just plot the average fluorescent intensity along the line for several sample for each condition.
How can I show that there is a statistical difference? What analysis (preferably free)?
Hi, I would use image J and create a macro which would roughly do this:
1. go along the line you draw starting from the minimum intensity moving towards maximum intensity (to make sure it goes from the outside of the cell, where the intensity seems lower even if the protein is membrane-localised)
2. Compare the current pixel value and the next pixel value to get the intensity gradient and go over the maximum gradient and stop where the gradient decreases below a set threshold. This will find the edge of the cell (if membrane bound, the gradient will get negative thereafter, if cytosolic, the gradient will became close to 0)
3. Calculate average intensity in a narrow band around the edge found in the previous step
4. Calculate average intensity in a broad band inside the cell (starting some distance from the found edge and continuing till the end of the line)
5. Calculate the ratio of those two average intensities (edge/interior). The ratio is expected to be significantly higher for membrane bound case)
6. If you do this for many images you'll get a set of ratios and calculate the mean and SD
7. Then you just compare your two samples (mean and SD for each) to see if they are significantly different.
Note, if you are sure that the edge is always in the middle of the line, you can even skip the gradient search and go directly to step 3, using the same portion of the line for each image.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Graphs