Hi Virginie,

companies that produce antibodies by immunization with peptide conjugates often offer so-called control peptides (same peptide sequence as for immunization but without carrier protein: BSA/KLH). You are supposed to add an excess amount of this peptide to your immunoassay, e.g. Western blot, and thus outcompete the binding of the antibody to its antigen (all antibodies bind to control peptide, cannot bind to the protein band on your blot anymore, band in the blot will disappear). Often that indeed will work. However, by that experiment you simply demonstrate that the antibody can bind to the control peptide. It is NO proof of specificity of the antibody, as it may still bind the wrong protein, as has been shown e.g. in this paper: Jositsch et al., Naunyn-Schmiedeberg's Archives of Pharmacology 379, 389-95 (2009).

To really prove the specificity of an antibody, I would tryone of the following options: i) compare wild-type and knock-out tissue.

ii) If this is not available, do a transfection with a plasmid encoding your gene of interest: only in the transfected lysate you should see a band of the right size in Western blot (or, if present in the cells used for transfection, the band should become much stronger upon transfection).

iii) Alternatively, you can knock down expression of a protein using siRNA: when on your Western blot the band of the correct size disappears or becomes fainter, you probably can regard your primary antibody as specific.