To test the strength of RNAi by feeding using the above optimized protocol, we compared it to RNAi by injection. We first tested a set of 14 known maternal-effect embryonic lethal genes (gpb-1, par-1, par-2, par-3, par-6, cyk-1, skn-1, dnc-1, bir-1, pal-1, dif-1, plk-1, dhc-1, and mex-3) to compare the lethality obtained with both methods (see [16,17,18] for a review of maternal-effect genes). All genes tested were 100% embryonic lethal by both feeding and injection except for par-3, which resulted in 97% dead embryos by feeding but 100% by injection, and cyk-1, which resulted in 55% dead embryos by feeding but 100% by injection (Table 3). For some genes, for example par-1, RNAi by feeding (n = 32 fed worms) and by injection (n = 9) both resulted in 100% embryonic lethality of the progeny in all cases. Whereas RNAi of par-3 resulted in 100% embryonic lethality for all injected worms (n = 9), worms subjected to RNAi of par-3 by feeding resulted in 100% dead embryos in 16/24 cases but lower levels of lethality (on average 88%) in 8/24
