Hi,

I think the second setup you suggest is preferred and you could even consider swapping the dyes for all groups too:

Gel 1: Cy2: pool, Cy3: group 1 (control), Cy5: group 2

Gel 2: Cy2: pool, Cy3: group 3, Cy5: group 1 (control)

Gel 3: Cy2: pool, Cy3: group 2, Cy5: group 3

That is, If you have sufficient sample and dye available, of course.

Hope this helps!

I agree with Eef Dirksen. Make sure you do not run the same label with the same condition on all the gels. There could be a bias in one of the labels, e.g. Cy5 binding slightly better to a certain peptide, which you will then mis-identify as up-regulated in the condition that you always labelled with Cy5. You don't want that happening. Instead, try labeling each condition with a different label in each gel.

This is referred to as "label swapping" and it's standard procedure for gel-based or gel-free proteomics (or any big-data experiment involving labels, really). Here's a publication that shows that label-swapping can definitely help you get more biologically relevant data: Article Effective correction of experimental errors in quantitative ...

Ok thank you, we always use the dye swap method by the way. Still i do not understand why is the second approach preferred..

We use Cy3 and Cy5. Most people stick with just two stains now (there is a Karp and Lilley paper that provides reasoning).

We run a standard in Cy3 and our samples in Cy5. We run large numbers of gels and use alignment to correct across gels. We don't usually dye swap as we like to keep everything run under the same conditions. A labelling bias between sample and standard will be present across all conditions as a fixed offset in our case and is less of a concern than it would be in smaller experiments.