Hi Dan,
Randomizing pixels is a worthwhile test.
With the Manders M1 and M2 coefficients if each fluorophore occupied half the ROI then you would expect around 25% of the area to show the presence of both fluorophores.
For the Pearson correlation coefficient I would expect randomization to produce a correlation close to zero - have you or anyone else ever found appreciable correlations after randomization ?
-----Original Message-----
From: ImageJ Interest Group [mailto:
[hidden email]] On Behalf Of Daniel White
Sent: den 12 augusti 2014 13:43
To:
[hidden email]
Subject: Coloc2
Hi sylwia. You really need to read the documentation we wrote for the coloc
2 plugin very carefully and also the papers hat are referenced there.
Fiji.sc/Coloc_2
PSF. Point spread function. The image of a point source of light is not a point. This is explained in the Costes paper The Costes randomization test is needed to tell you if the Pearson's and Mander coefficients you measure are higher than a randomized version of your image. If yes then good. If not then you have no better than random colocalisation and should not infer any biological phenomenon.
Read all the docs. Read them again and the papers. Re read the docs. Digest then new information. Then ask us more questions. It tool me months to understand ....
Best
Dan
--
ImageJ mailing list:
http://imagej.nih.gov/ij/list.html--
ImageJ mailing list:
http://imagej.nih.gov/ij/list.html
Locked
