Looking for a better method for quantify protein intensity in a fluorescence image

Image Analysis
, ,
#1

Hello everyone

I’m trying to quantify and compare the intensity of my interested proteins from mutant cells and WT cells. I took pictures with a 40x objective lens and cropped several cells per picture (see sample image). For quantifying the signal in each cell, I created ROI by freehand selections in the green (GFP) channel and applied it in the red channel to perform the measurement. Then I did statistical analysis for obtained mean value or integrated density.
sample.tif (631.2 KB)

I have several questions:

  1. which value should I use for statistical analysis? mean or intDen?

  2. should I use a 20x objective lens to obtain more cells per picture to increase efficiency? Usually, I have many mutant groups and I want to quantify more cells in one group. So far, I just take less than 100 cells from one group.

  3. If I take picture with more cells under the 20x lens, is there any automatic method to trace the outline of the cell? I want to create the ROI of the cells which expressed GFP and measure the intensity of each cell by dividing the total red signal intensity by the number of cells. I know I can create a mask after thresholding, but unfortunately, I don’t want to measure the dividing cells or the cells with multinuclear or on the edge.

Do you have any suggestions?
Thanks.

#2

Hi JZZ3090,

  1. Most people use Mean grey values but IntDensity is useful if the area of WT and mutant cells is different. Only you know your cells.

  2. It depends on the microscope and lenses you have: if your microscope is semi-automated, and can take many pictures easily, use the 40x. If not, and the 20x lens is good enough to image the filamentous structures you are looking at, the 20x should be fine. This means that you will be able to count the amount of red fluorescence in a cell, but will not have resolution enough to look at individual strands.
    I see from the image metadata that you have used an olympus confocal with a 40x (N/A 0.9) air (?) lens to take this picture. Why not take pictures with a widefield microscope instead? The advantage is that you’ll see whole cells (not just a slice, so mean intensity will not be biaised since you want to measure a whole cell) and you’ll take your pictures much faster.

  3. I’d suggest that you learn to use CellProfiler -it’s designed exactly for this kind of analysis.

Sincerely,

Matthieu

1 Like
#3

Thanks for your suggestions!
I don’t need more details in a cell so 20x may be ok.
I did confuse about where I should focus when I used a confocal microscope and I’ll consider the widefield microscope.