While RNAi has proven a tremendously valuable approach in studying the role that genes and proteins play in a myriad of biological processes and diseases, it has three main limitations that gene editing is able to overcome.
Identifying an siRNA or shRNA that can completely ablate expression of a target transcript can in some cases be challenging. This can be an issue if the residual expression is sufficient to mask any loss-of-function related phenotypes, leading to false negatives.
Gene editing enables knockout of the gene at the genomic level, completely preventing expression of functional transcripts, and overcoming masking of phenotypes.
(An exception to this is instances where a gene has alternative splice variants, or secondary start codons. In these cases rather than relying on frameshift mutations to knock a gene out, a larger deletion event may be required, or the frameshift mutation should be introduced into a region of the gene encoding a functional domain.)
The potential for off-target effects with RNAi is well documented - where the processed siRNA or shRNA leads to downregulation of an unintended target which in turn results in false positives.
With gene editing, off target risk can be mitigated by creating independent clones in the case of cell lines, or by breeding out in the case of animal models. Further to this, predicted off-target sites can be sequenced to look for evidence of off-target modification.
Especially at the level of screening, RNAi has exhibited issues with reproducibility, which could in part be related to issues with off-target effects. This is highlighted by the example of three screens performed to identify HIV host factors, the overlap between which was three genes.
Evidence suggests CRISPR based screens are significantly more reproducible. We have for example, been able to recapitulate the results of a number of published CRISPR screens as part of validating the technology for our CRISPR screening service.