Hi David:
One solution for this is to use other publish methods of FRET normalizations or conversions:
1. FRET/CFP is the one you are using
2. FRETN = FRET/(CFP*YFP)
3. NFret = FRET/sqrt(CFP*YFP)
4. E%
5. Distance
I myself have a similar problem solving for the values are simple with ImageJ but after solving image arithmetic I noticed that I get extremely large pixel intensities. The reason is that the numerator (FRET pixel value) is divided over small numbered denominator or zero background (CFP or CFP*YFP intensities). This happens since backgrounds have few or relatively no signal. The image division yields a divide by zero resulting in infinite (the high image pixel). Upon inquiry with a few people the solution is to assign pixel intensities above e.g. an arbitrary level e.g. 200 (high pixel intensity are close to 65535 for 32bit image).
I do not know how to set intensities above a level to zero. Maybe you have an answer to that.
-Nguyen
Postdoctoral Researcher
UNC Chapel Hill
Chapel Hill, NC
--
ImageJ mailing list:
http://imagej.nih.gov/ij/list.html
Locked
