Measuring protein expression intensity

[Demilux]

Hi list,

I'm quite new to ImageJ, so I have no real clue how I should approach what I need to do.

I'm aware that quantifying is relative when it comes to bright field microscopy, but I still want to extract some relative numbers from my photographs. I am to "measure" the intensity of the expression of a protein in certain ROIs. It was suggested to me to use histograms and extract values from them as a means of relative quantification.

However, I have several problems:
1) I'm not sure this is the best way to get relative values for protein expression. If that's so, what method would you suggest?
2) If I am to use histograms, how exactly do I get values from them?
3) The backgrounds in my photos are not always pure-white; how can I adjust that without altering the non-background?
4) Should I use the background as point zero for my histograms or the tissue which shows no expression of this protein?

Here's a sample picture of a plate I'm dealing with (http://postimg.org/image/84gc6ijhb/).

Any help would be appreciated!

Thanks ahead,
Dlx

[Gabriel Landini]

On Tuesday 04 Mar 2014 13:59:34 Demilux wrote:
> I'm aware that quantifying is relative when it comes to bright field
> microscopy, but I still want to extract some relative numbers from my
> photographs. I am to "measure" the intensity of the expression of a protein
> in certain ROIs. It was suggested to me to use histograms and extract values
> from them as a means of relative quantification.

It is not "relative", it is "unreliable". That's different.
So no matter how reproducible the imaging part is, the optical density of the
ROIs is not a true representation of the protein expression.

Please see the mail archives in relation to quantifying immunohistochemistry.


Regards

Gabriel

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

[Demilux]

Thanks for the reply.

What bothers me is that everyone is suggesting the colour deconvolution plugin, which is used when one needs to separate colours/stains in staining (from what I got). What I need is just to quantify the intensity of a BW image, and I can't seem to use the plugin for that (maybe it's possible, but I don't know how to properly use the plugin). If this is possible, would someone be kind enough to explain how?

Another thing which I fail to understand (I'm new at this, don't take it hard on me) is why exactly "quantifying" bright field IHC staining is considered unreliable. The brighter the colour, the less protein there is expressed, and vice versa. So, why is it a bad idea to assign relative numbers to the intensities (0 = black = max protein expression; 255 = white = no protein)? I would use the same light intensity settings on the microscope/camera for each plate and then compare them one to another. As I said, someone suggested using histograms. But, if this is such a bad idea (?), do you have any other suggestions?

Thanks ahead

[Gabriel Landini]

On Wednesday 05 Mar 2014 10:17:34 Demilux wrote:
> Another thing which I fail to understand (I'm new at this, don't take it
> hard on me) is why exactly "quantifying" bright field IHC staining is
> considered unreliable.

In the "note" section in red, are mentioned some of the problems:
http://www.mecourse.com/landinig/software/cdeconv/cdeconv.html

Regards

Gabriel

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