Published and unpublished off-target RNAi effects?

I know there have been computational studies regarding the potential for off-target effects of dsRNA treatment in worms, but does anyone know if this has been tested experimentally? And, a slightly different question, how often do people find that RNAi of their gene of interest results in a phenotype different than the known null phenotype?

we tested a collagen dpy-13 gene. dpy-13(e458) is likely a null, with a length half of N2.
If we RNAi ergo-1(-), or, eri-1(-), or eri-1(-);dpy-13(e458) animals with dpy-13 targeting the deleted region, worms elicit a worm-ball-like phenotype that is dramatically more severe than dpy-13(e458). We called it superdumpy. Nuclear RNAi (Nrde) likely likely plays a major role here.



I am doing a RNAi module in a Genetics lab and am interested in using a construct that causes worms to be Dpy. If your construct makes nicely Dpy animals, would it be possible for you to send your bacteria to me? I have been using myo-3 RNAi, but the students have trouble identifying the paralyzed worms. I am hoping a viable but clear phenotype would work better. Also, do you know how strong the phenotype is in N2 (vs. eri-1)?

Thanks, Janet Duerr (duerr at ohio dot edu)