Question regarding fluorescence colocalization

[Daniel White]

Hi Sivram,

Jeremy is dead right, colocalization  a potential minefield.
To help everyone navigate the pitfalls,
we wrote a new imageJ plugin for intensity corellation based methods, to
help standardize things and squash software bugs, Coloc_2
and a bunch of documentation to go with it, practical and theoretical tips.

Here are links to our Fiji colocalization plugin, Coloc_2   (it's part of
Fiji by default - nothing extras to install), and the documentation to go
with it.
http://imagej.net/Coloc_2
http://imagej.net/Colocalization_Analysis
http://imagej.net/Colocalization_-_hardware_setup_and_image_acquisition
http://imagej.net/Chromatic_shift_origins_measurement_and_correction
http://imagej.net/Category:Colocalization

Teaching material
https://www.biodip.de/w/images/f/fa/QuantitativeColocAnalysis-10-2011.pdf

Do read the Manders and Costes papers that explain what the plugin does.

(Object based methods is another story, its a segmentation problem,
followed by a colocalization definition problem. )

Always define a biologically meaningful region of interest to analyze.
Rarely the whole image.
 Deconvolve it to restore proper contrast of small features and suppress
noise. Noise kills correlations.
Subtract background and detector bias/offset.
Watch out for colour channel excitation crosstalk and emission bleed
through, especially dapi into green, and red in to far red.

Explicitly state the spatial resolution that you are measuring correlations
at.
The whole universe is one voxel - everything is colocalized.
2 particles cant have the same quantum numbers - nothing is colocalized.
We are measuring at some spatial scale  in between... but where?

The higher resolution imaging you do, the less colocalization you will
generally see.

Show scatter plots rather than colour merge images... but if you must do
colour merges, only do 2 colour at a time, and give the colour blind a
chance: use green-magenta or red-cyan.

Looking for yellow is a big No-No...
See
http://imagej.net/Spirals


Best regards,


From On Behalf Of sivaram mylavarapu
Sent: den 2 november 2016 15:13
To: [hidden email]
Subject: Question regarding fluorescence colocalization

***** To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy Post images on
http://www.imgur.com and include the link in your posting. *****
Dear colleagues,

I am looking for an authoritative review(s) on the principles and best
practices to be followed for measuring immunofluorescence co-localization,
including obtaining quantitative information like the Pearson's
coefficient. It would be very useful if the most common pitfalls to be
avoided were also dwelt upon on in the articles. I would appreciate if you
could point me to such literature.

Thanks much!

Sivaram.

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html