"

From 2010.igem.org

"It has become dangerously acceptable to conclude that proteins interact directly provided that (i) the interaction is biologically plausible; (ii) the proteins are co-expressed; and (iii) an interaction is demonstrated by glutathione-S-transferase (GST)-pulldowns and/or co-immunoprecipitations (co-IPs)"... "The most obvious caveat with co-IP data is that a positive result does not imply a direct interaction between two proteins – binding could be mediated by other partners."..."In addition, a widespread misconception is that co-IPs from extracts provide in vivo evidence for the existence of an interaction. This is not accurate, particularly when the experiment is carried out using overexpressed proteins in cell lines."...""

""

"Protein interactions: is seeing believing? J. P. Mackay et al. 2007"