Bug fixed in ImageJ/Fiji Coloclization Threshold plugin

[Daniel James White]

Dear All Colocalization fans,

After a strong hint from the Perkin Elmer / Volocity developers,
and during the design of the next generation Colocalization tool for imageJ,
we believe we have located and fixed a bug in the imageJ plugin called
Colocalization Threshold.

This plugin comes from the MBF-ImageJ collection (Tony Collins and Wayne Rasband are cited there as the authors)
and is included in the "Fiji" imageJ distribution (in a modified form containing some enhancements)
http://pacific.mpi-cbg.de

The plugin implements the auto thresholding method using Pearson's Correlation Coefficient,
as described by Costes et al 2004.
It also provides many statistics and measurements and a scatterplot, as output.

One important statistic it measures is the Pearson's correlation coefficient above thresholds: Rcoloc.
see
http://www.macbiophotonics.ca/imagej/colour_analysis.htm#6.3 Colocalisation Threshold

This was calculated incorrectly, but we have fixed it.
This fix will hopefully soon be in the Fiji updater, so you can all get it easily.

What it means for users is that Rcoloc figures were lower than they should have been, according to the description of Rcoloc at
http://www.macbiophotonics.ca/imagej/colour_analysis.htm#6.3 Colocalisation Threshold

Again, thanks go to Andrew and Steve from Perkin Elmer  - makers of Volocity,
for bringing this problem to our attention so we could know to fix it.

cheers

Dan White




Details for those interested:

In the MBF version source code
line 695 is commented out,
but we think this is an error,
since indeed the number of pixels to be taken into account in the following calculation of Rcoloc
should be only those above threshold (and in the ROI/mask).

We have just pushed a fixed version of the source code for this plugin to the Fiji git repository contrib branch,
and you can see the fix and the commit comment there
(and also other fixes we made recently)
We have made the use of the Ncoloc member variable in the Rcoloc calculation explicit,
as opposed to using the member variable N.

I guess this was a copy paste bug or similar...since Pearson's calculation seems to be done three times over
in different places in the plugin source code.

Please let us know if you think we have made a mistake here.

If Tony agrees, he might like to also update the version in the MBF plugins package distribution.

Dr. Daniel James White BSc. (Hons.) PhD
Senior Microscopist / Image Visualisation, Processing and Analysis
Light Microscopy and Image Processing Facilities
Max Planck Institute of Molecular Cell Biology and Genetics
Pfotenhauerstrasse 108
01307 DRESDEN
Germany

[phone] (German Mobile)
[phone] (Work phone at MPI-CBG)
[phone] (Fax MPI-CBG LMF)

http://www.bioimagexd.net  BioImageXD
http://pacific.mpi-cbg.de                Fiji -  is just ImageJ (Batteries Included)
http://www.chalkie.org.uk                Dan's Homepages
https://ifn.mpi-cbg.de  Dresden Imaging Facility Network
dan (at) chalkie.org.uk
( white (at) mpi-cbg.de )