Today's paper is: 

Legrain, P. & Rain, J.-C. Twenty years of protein interaction studies for biological function deciphering. Journal of proteomics (2014). doi:10.1016/j.jprot.2014.03.038.

I keep changing the reference format. This really shouldn't bother anyone but me.

The take-home message: There's a lot of protein interaction data out there! Contrary to popular belief, most of it isn't just false positives. Rather, most of this data reveals actual biological complexity. Proteins may just interact with more binding partners than we originally thought.

A few awkward points:

• Both of the authors are experts in the field but are also employed by Hybrigenetics, a company providing protein interaction screen services. They don't advertise the company's services specifically so I suppose it isn't really a conflict of interest. I still grow concerned about such things.
• The title sounds odd. Couldn't it have been "Twenty years of biological function deciphering by protein interaction studies"? I think it's the word "deciphering", mostly.
• The entire review sounds a bit strange, actually. I assume it's a result of English as a second language. It does make some conclusions hard to understand, i.e. "Almost 2000 different proteins were analyzed over six time points, covering four orders of magnitude in terms of protein abundance. In those papers, the aim of the purification process is no more the isolation of a protein complex but just a way to zoom in a specific part of the proteome", which sounds dismissive.
• More specific references to recent interactome-dependent functional studies would have been nice. As they mention, there have been thousands, but it continues to be an active research venue. It looks like the authors just forgot to add some references in other sections: i.e., "..in different cellular contexts (for review on affinity-purification coupled to mass spectrometry, see)".

These guys made an early protein interactome of Helicobacter pylori so I can't complain too much.