How to know if the signal of western blot is saturated ? - (Jun/04/2013 )
I heard that we can't semi-quantify proteins if the picture of the western blot is saturated. But how do we know it's saturated ? I know that the capture software display saturated pixels in red but is that all ? Since the file format of the capture software is unusable I usually take a screenshot or copy-paste the picture in Paint and I save it in .bmp format. The problem is when I then open it with imagej and check the intensity of my pixels and I see "0" or "1" for most of the pixels located in the middle of the bands. It looks like it is saturated to me.
So the question is : is it really saturated or does imagej index darkest pixels to 0 and brightest to 255 ? If it is saturated where's the problem since the capture software doesn't mark any pixel as saturated ?
Saturation basically means that the signal has reached a maximum for that particular pixel - hence you can't get any more information out of it than it already contains. Taking a screen shot or other low resolution format is not a good way to do this. There should be an export or save as function that will allow you to get the full image in a transferable format.
The brightest and darkest calls you are making are subjective - are you using a positive (black bands on a white background) or negative image (white bands on a black background)? Is it grey-scale, colour, binary, how many bits?
Which of the imageJ tools are you using for the analysis?
Yes the picture is greyscale, bands are black and background is white and grey.
With imageJ (Fiji last version) you can put your cursor on the picture and see the value of intensity of the pixel beneath. In my case when I check bands I see that most of pixels in the middle of bands have a value 0 (= black = darkest). But another hypothesis is the pixel has a float value and imageJ only display the rounded value.
OK, ideally those pixels would have a positive reading - a few saturated pixels won't hurt too much, but if the majority of the band demonstrates those values then there could be a problem. You can use the wand tool to link pixels of similar value, that will give you an estimate of the proportion of the band that is at the 0 point.
After trying the export option on the software I think the values are scaled from 0 to 255 and rounded, saturation only occurs when the softarware highlight pixels in red.