at the mRNA level, "good" knockdown is >70%, ideally >80% - in cells (measured by qRT-PCR). at the protein level its always less, and in general 50-70% is good, tough to accomplish more. it depends on many things including protein half life

Usually, 50% knockdown is considered acceptable. To be sure that you have a potent knockdown, kindly confirm it with Western blot as well.

There is no magic target value. The better the knockdown at the protein level, the more likely you will see a phenotype. But how you define success depends on what your goal is. For example, if you want to study the function of NRF2, you should be checking known NRF2 target genes for changes in expression as a function of knockdown. If you can see 2x or larger fold changes in target genes, then you might interpret your knockdown expt as a success. If you're trying to validate an anti-NRF2 antibody, then a 55% knockdown may not be convincing enough.

Plenty of examples in the literature where 50% knockdown of a protein yields a phenotype. Even more where >=90% knockdown does not (perhaps a small amount of protein goes a long way or other proteins are functionally redundant). The answer is entirely empirical and system-dependent (function of GOI, cell-type, assayed phenotype, etc).

I am personally always skeptical when I see a minor knockdown efficiency such as 50% on the protein level and a "big phenotype" is claimed. Anyway, I think it is absolutely essential that reconstitution of protein expression in the context of the knockdown (using an siRNA/shRNA-insensitive wildtype protein) is performed and that it is demonstrated that this reconstitution fully rescues the phenotype.

at the mRNA level, "good" knockdown is >70%, ideally >80% - in cells (measured by qRT-PCR). at the protein level its always less, and in general 50-70% is good, tough to accomplish more. it depends on many things including protein half life

Most people (and the editors who will reject your paper) don't like to see less than 50% reduction. However, some genes are hard to knock down, you can waste a lot of time trying to get better knock downs. If there is no published knock down reference on that gene, I will go to where I see the phenotype i'm looking for and where I can rescue it. I guess you can always argue with the editors based on that.

completely agree with Daniel- we screened and selected top performing siRNAs for 100's of genes (>90% knockdown at the mRNA level); there was a handful of targets for which we could not select a single siRNA that would induce more than 40-60% knockdown, despite we tried many designs. its not clear what is the problem- localization, structure, inaccessibility due to proteins bound, etc. but keep in mind- there are some very rare targets like that