image analyis -need some help - analysis of fluorescent microscopic images (Oct/21/2006 )

Hi,
I need to quantify the fluorophore uptake by different cells during a given timeframe. I've been looking onto the ImageJ software, which seems perfect to the task, and I think, I even know what I want to do, yet I'm still completely lost. I looked into different online guides, but most of the terminology is completely foreign for me.
I think I should quantify the area of the cells AND the intensity of the pixels inside of the cells. I need to take it in account that the cells are not uniformly spread on the different petri-dishes, and I might have more or less cells visible in different experiments -so I guess average intensity would not do.
I thought about transforming the images into 3D plots and i ntegrating their area, but I got stuck at the 3d Plot part. (They look cool, though.)
OK, so I'm really new in this field. Even if I find an answer I'm not perfectly sure if this particular answer would be acceptable in a peer reviewed paper. Please help... How do the gurus of fluorescent microscopy do that?

Hi spongya,

Image analysis may be simple or extremely diffuicult, depending on your pictures. I have recently started working on image analysis, too, and here is what I've learned:

In principle, the computer first divides the picture into areas of particular interest, i.e. segmentation or "finding particles". In your case the segmentation process should find the borders between cells. This is the most difficult part, even for the human eye. Distinguishing between nuclei, for example, is simple, whereas distingwishing between labeled membranes is highly difficult.

Once you have a correctly segmented image with clear boundaries you can get statistics for each region. Total fluoresence, for example, will be the sum of pixel intensities. The area will be related to the number of pixels (although this must be treated with caution because particles with strong fluorescence appear larger due to off-focus effects), and the mean fluorescence is the mean pixel intensity.

ImageJ may or may not be the best for your purposes, depeding of the complexity of the images and the number of pictues you need to analyze. There are other programs that do a better job. I have heard of the commercial program Metamorph, and there is also a free Matlab-based program called CellProfiler at http://www.cellprofiler.org/.

Good luck!