Technologies are democratizing insect genetics and it seems as if everyone is thinking about making ‘knockouts’. This is a good thing but as Housden et al. nicely explain in a recent paper in Nature Reveiws Genetics explain:
“although LOF (loss of function) approaches may initially all seem to achieve the same effect, that is, disruption of gene function, there are considerable conceptual differences among them that can lead to substantially different outcomes.”
The nuanced differences to which Housden et al. refers are worth appreciating by anyone attempting to create loss of function mutations/phenotypes and/or interpret the results of such efforts. For those not working with model systems for which there has been a strong tradition of using genetic approaches some of these differences may be underappreciated.
The loss of function approaches discussed by Housden et al. include chemical mutagens, transposons, RNAi, morpholinos, Genome editing, CRISPRa (activators) and CRISPRi (inhibitors), Degrons and Small-molecule inhibitors.
One particularly interesting point Housden et al. make is that the results from experiments using different loss of function approaches can quite often be inconsistent. Sometimes this is due to easily imagined technical differences such as differential efficiencies. There may be differences in when the loss of function technology was delivered, resulting in different outcomes. There may be off target effects that are of more or less importance depending on the technology.
But, interesting, as Housden et al. discuss, different outcomes may have their origins in biology. Here is one of the examples they use. In zebrafish a knockout mutation in egfl7 created using TALENS had a less severe phenotype than that seen following two knockdown approaches – morpholinos and CRISPRi. Most might have interpreted those results as reflecting off-target effects associated with the knockdown technologies employed. But in this case, this interpretation would be wrong. Instead, the strong effects of the TALEN knockout were compensated for by elevated expression of other genes. These compensatory changes in the expression of other genes did not occur following mere down regulation of the gene.
The authors review a number of examples that further illustrate how confounding results can reflect real and important biology. Very interesting. Very enlightening.
These examples, while illuminating, have the net effect of reducing the confidence in ones interpretation of the results of loss of function experiments (a potentially healthy effect). But how to mitigate this reduced confidence? How to increase confidence? When studying single genes, for example, it might be prudent to employ multiple loss of function methods. At the very least it is worth carefully considering the approach to be used and to understand fully what one can and cannot learn from the approach.
The paper by Housden et al. should be ‘required reading’ in laboratories working with loss of function technologies. As the authors conclude, although seemingly straightforward, loss of function approaches and their interpretation are complex and failure to appreciate this fact has and will result in “false discovery and inaccurate conclusions”.
Loss-of-function genetic tools for animal models: cross-species and cross-platform differences. Benjamin E. Housden, Matthias Muhar, Matthew Gemberling, Charles A. Gersbach, Didier Y. R. Stainier, Geraldine Seydoux, Stephanie E. Mohr, Johannes Zuber & Norbert Perrimon Nature Reviews Genetics (2016) doi:10.1038/nrg.2016.118