Of the 168 genes, 123 were determined to be embryonic lethal by either method, of which feeding detected 97% and injection 91%. The embryonic lethality was of equal penetrance using either method for 77% of these genes, was stronger by feeding for 15% of the genes, and was stronger by injection for 8% of the genes. Of 52 genes giving a sterile phenotype by either method, 96% were identified by feeding and 48% by injection; 44% were detected by both methods, 52% only by feeding, and 4% only by injection. And, finally, of 154 genes giving a post-embryonic phenotype by either method, 96% were identified by feeding and 64% by injection; 60% were detected by both methods, 36% only by feeding, and 4% only by injection. In addition to the above, we also injected 30 random genes that gave no phenotype by feeding, and none gave a detectable phenotype by injection (data not shown). Thus, from this expanded data set, we conclude that RNAi by feeding is roughly equivalent to injection for detecting embryonic lethality, and is somewhat better for detecting genes causing sterility or other post-embryonic phenotypes.
