I’m looking for previous articles in which authors were able to demonstrate a phenomenon based on overexpression experiments, but were not able to obtain corresponding (i.e. reverse) results based on silencing experiments (eg. via siRNA). I’d be very grateful for any reference to relevant articles.
If there is a second gene product that complements the activity of the targeted gene, a knockdown might have very little phenotypic effect even though the overexpression causes a phenotype. I don’t have a citation handy for that. Did you confirm your knockdown immunochemically?
Jon’s got a point there. Additional copies of the same gene, a similar gene with the same function, or a second gene product would not show the same effect as if you were targeting a single novel gene. The effectiveness of your knockdown (silencing with siRNA) will determine whether you get the desired effect. Did you use the exact same approach as that cited in the other paper? The transient nature of siRNA and the extent of knockdown can vary widely, especially if you did not use the same construct and conditions as that citied in the other paper. Protein turnover and the protein half-life may determine whether the efficacy of knockdown is sufficient to get the expected result. You may have to make stable transfectants using a plasmid construct expressing a shRNA to your gene of interest. Ultimately, knockout mice might be an option if this is something that has to be validated. Did you check mRNA and protein levels for the gene target?