Search Vadim's Blog

Sunday, December 10, 2006

To do or not to do

A number of researchers contacted CoLocalization Research Software Support lately and said they tried either CoLocalizer Pro or Express versions of colocalziation analysis software and the results of colocalization coefficients were not what they expected. They wondered why is this so and what should they do to obtain more suitable calculations. After asking them few questions about the way they prepared samples and how they used confocal microscopes, it became clear that their images were simply unusable for quantitation purposes. For these images, even calculations which seem correct will no doubts be erroneous. We put together some simple guidelines for researchers using confocal microscopy and doing quantitative colocalziation analysis so that they can avoid mistakes in future. Here they are:

Sample preparation:
- Make sure that the used antibodies are specific and do not cross-react.
- Choose fluorophores with well-separated excitation and emission spectra.
- Use the same mounting medium for the samples you will be comparing.
- Try to avoid using anti-fading reagents, as they can increase background fluorescence.

Microscope set-up:
- Use plan apochromatic lenses to reduce chromatic shift.
- Use optimized emission filters to maximize emission collection while avoiding bleed-through effect.
- Use the same objective lens for the samples you will be comparing.
- Make sure that the size of microscope pinhole is properly set.

Image acquisition and handling:
- Acquire images sequentially to minimize bleed-through effect.
- Do not acquire too bright and too contrast images, as it will result in their saturation.
- Do not manipulate images in graphics-editing programs, as it may ruin usability of your images for quantitation.
- Do not resave files in any other format than TIFF. It may cause loss of original image data needed to perform calculations.

If you have not followed one of these guidelines, just don`t use your images for calculating colocalization. You will not get reliable results.

No comments: