I work on the mosquito Anopheles gambiae and generally, we use dsRNA about 250bp in length, however I have used dsRNA as short as ~120bp and as high as 600bp. I've found that the relative efficacy of any one dsRNA has depended more upon the specific dsRNA and target gene rather than the length. I've had variations in knock-down when targeting the same gene with a number of dsRNAs all between 150-250bp, yet the greatest variation is between target genes. It also seems that efficiency is dependent on the initial relative quantity of the target gene which makes it a little difficult to assess if you are not looking at a housekeeping gene. As a general rule, when I have targeted the same gene (with ~250bp dsRNA) as other groups (using siRNA) I've managed to replicate the level of knock-down, which has itself varied between target genes. I would also like to add that the concentration of dsRNA is of utmost importance. Although knockdown efficiency reaches a plateu as the concentration increases, the plateu tends to vary by target gene. This is all comparing dsRNA delivered by thoracic injection, there is some progress with other methods of delivery (larval feeding, laval rehydration, ectopic siRNA expression from plasmid etc.) yet they seem to be difficult to replicate. In fact, a previous discussion has verified that quite a few groups have struggled to replicate these different delivery methods, at least in the mosquito. I appologise that I haven't got any nice experimental comparisons but there are so many factors to consider - time after delivery, dsRNA specifics, dsRNA length, delivery method, endogenous expression level, specific target gene and even gene expression quantification method.

Thank you very much Andrew. I have silenced genes in adult sand flies in the past, but I usually design primers to cover most of my putatively expressed gene of interest. I have used 600-700 bp long dsRNA in the past and delivered it by nanoject-micro-injection and it worked. I am now trying to silence different genes in sand fly larvae using alternative delivery methods (feeding, soaking, etc). My first feeding approach was not successful, but I have been using a low concentration of dsRNA. I am repeating my experiments with a higher dose but I was wondering if I might be worth a thought re-designing my primer to cover a smaller region. I thought about it after speaking to Brazilian colleagues that work with Rhodnius prolixus and short dsRNAs(100-200 bp). Don't apologize for not having any experimental comparisons, your answer has been really helpful!

I've had just one go at the feeding method and have heard of others who have have trouble with both feeding and soaking. However one of my main problems was that the dsRNA/chitosan/food mix was insufficient for the larvae to develop as normal. How about injections of an siRNA expression vector (~23bp siRNA) into the embryos? Although embryo microinjections are notoriously tricky. I would definitely start off silencing a gene you have already silenced before when trying these methods. I'm guessing you are looking at the alternative methods because you want to knock-down a gene earlier on in development? If it is because dsRNA injected directly into the thorax was successful for some genes but not others, it could be because of poor tissue penetrance or a couple of other things such as highly regulated gene expression. I've had trouble with knock-down for genes which are very temporarily induced, which could be because the RNAi doesn't have enough time to act, or because the level of induction completely overshadows any knockdown. In the end I didn't find a way around silencing those genes.

greetings to all

i am also working in this area and my concern is that, if the mode of administration is injection then the length of dsRNA should affect the efficiency of RNAi??

Length is a concern in cell autonomous RNAi or systemic RNAi or both. i think that ultimately all the dsRNA gets chopped by dicer in the cells. please correct me if i m wrong. however in case of oral feeding i think that internalisation of dsRNA molecule is an important factor. has anyone tried soaking for insects like aphids..??

Well systemic RNAi as it exists in plants, i.e. spreading and amplifying, doesn't exist in most insects and definitely not mosquitoes or fruit flies, although it works well in Tribolium. So for most insects, I think that length could only be a concern if specificity issues arise because, potentially, some of the chopped up fragments have perfect matches to other RNAs. This would risk would obviously increase as the dsRNA length increases. Then again, single siRNAs may have very variable efficacy which could be averaged out in a longer dsRNA. As well as off-targets, long dsRNA molecules may elicit some defense mechanisms which could give strong phenotypes capable of masking the true RNAi phenotype. However you would imagine that a good control would be enough to limit the concern...unless secondary structures are added into the mix. As for the soaking of dsRNA, I'm not sure if I've read anything about length of the dsRNA and efficacy, however, the amount of dsRNA required for the method is very impracticable, especially given that it is unrefined.