Re: question regarding Colocalization Threshold in Fiji - Coloc2 (fwd)

[Daniel James White]

Hi Giovanna,

(Tom, we have a nasty bug to fix)

On Jul 29, 2011, at 2:06 PM, Johannes Schindelin wrote:

> In case you missed it...


Thanks Johannes! I did miss it!

Its best to email directly to the person you want as well as to the mailing list,
as sometimes i miss messages that are in the middle of the daily imageJ digest email.

Or maybe I didnt get the digest yet today...
anyway, thanks Johannes!

> ---------- Forwarded message ----------
> Date: Fri, 29 Jul 2011 03:00:46 -0700
> From: jo <[hidden email]>
> Reply-To: ImageJ Interest Group <[hidden email]>
> To: [hidden email]
> Subject: Re: question regarding Colocalization Threshold in Fiji - Coloc2
>
> Dear Daniel,
> I have two questions concerning Colocalization threshold and Coloc2 plugin
> which I´ve been testing lately to analyze my data.

OK
>
> I applied the Colocalization threshold plugin on 2 background-subtracted
> images for example one “positive” image where colocalization is
> expected/”visible” and one “negative” image with no colocalized signals.
> Then, I´ve tested the results with the Colocalization test plugin and all
> the p values were much lower than 0.

Huh? Hoe can p values be less than 0

Do you mean the randomised image Pearsons r values were all less than the
Pearsons r of the real images?

I guess so?

> I´ve also tried with Coloc2 and the results (the tM1 and tM2) were extremely
> different from those of Colocalization threshold. The thresholds were almost
> the same (Ch1 equal, Ch2 25 instead of 26) but where NO colocalization
> occurs: Colocalization threshold provides tM1 0,5388 tM2 0,5833 whereas
> Coloc2 Manders M1 (threshold): 0.836 Manders M2 (threshold): 0.786.
> I don´t see the colocalization but the Manders Coefficients from Coloc2 are
> quite high.

Today I alos noticed unusually high thresholded Manders numbers in
Coloc2 compared to Colcoalization threshold plugin.
Actually I got numbers higher than 1,

so we have found a bug!
We are looking in to it now. Hopefully its simple.

> I´d rather use the Colocalization threshold at this point. What do you think
> about?

Right now there seem to be cases where Coloc2 is giving wrong thresholded Manders1 and 2 values,
so at the moment the old plugin should do the right thing.

but watch out, the old plugin has some issues as well, and may also be buggy in other ways...
that why we made a new one.
>
>
> On the same pictures the Coloc2 plugin gives me an error:
> Warning! y-intercept high - The y-intercept of the auto threshold regression
> line is high. Maybe you should use a ROI. What is the reason for such a
> warning and how I can improve my analysis if I don´t want to use a ROI?

You must always restrict the analysis of colocalization to the biological structure you are
interesting in measuring. Its no good to analyse the whole image,
including large areas of background and non specific signals.
(the only time not to use an ROI, is when every single pixel of the image might contain interesting biology)

this is clearly stated in the Costes et al paper of 2004,
as linked to here
http://fiji.sc/wiki/index.php/Colocalization_Analysis

In coloc you you can use an normal imageJ ROI, however you loike to make it,
or a mask image, as the Coloc Test plugin can,
so it can be a 3D ROI in effect.

I very strongly recommend you read the Costes and Manders papers
linked to there so you understand what the maths is trying to do
and what kinds of images will cause it to fail. Its not bullet proof,
and you have to give it the right kind of food.

http://fiji.sc/wiki/index.php/Colocalization_Analysis
tells you many useful and important hints... read carefully, and digest, then read again.


> Many thanks for your work, I´ll really appreciate any comment/suggestion

Its a pleasure to be able to help you,
and we greatly appreciate your input.

How would you like the output of Coloc2 results to be standardized?
What should the PDF output look like

All opinions welcome.

cheers

Dan

> J
>
>
> --
> View this message in context: http://imagej.588099.n2.nabble.com/Re-question-regarding-Colocalization-Threshold-in-Fiji-Coloc2-tp6035902p6632945.html
> Sent from the ImageJ mailing list archive at Nabble.com.

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 )