In 1990, Richard Jorgensen's laboratory was trying to create petunias with a deeper purple hue. They reasoned that introduction of a transgene should up-regulate the production of pigment. Instead they achieved an unexpected result: Some flowers showed variegated pigmentation and others had sections with no pigment at all.

Rather than deepening the purple color, the transgene appeared to be inhibiting it. Jorgenson called this phenomenon co-suppression, and attributed it to a mutual inhibition between the endogenous and transgenic mRNA. Soon, other groups began reporting similar observations and showed that endogenous and transgenic mRNAs were being degraded after transcription had occurred. This phenomenon of co-suppression was termed post-transcriptional gene silencing (PTGS).

Evidence of homology-dependent gene silencing in organisms other than plants began to emerge. Similar observations, termed quelling, by Giuseppe Macino's laboratory were seen in the filamentous fungus Neurospora crassa when a transgene was introduced to increase the expression of an orange pigment in the fungus. Rather than turning orange, a third of the transformants actually became albino. Macino was able to show that the inhibition was due to quelling, not a change in the rate of transcription of the pigment gene.

RNAi, however, moved into the limelight after work in the nematode worm Caenorhabditis elegans. In 1995, Guo and Kemphues were using injected antisense RNA to suppress par-1 in C. elegans. They were surprised to find, however, that their negative control, sense RNA, also mediated suppression.  Finally, in 1998 Fire, Mello, and colleagues unequivocally demonstrated that double-stranded RNA (dsRNA) was able to direct degradation of homologous mRNA. Like Guo et al., Fire, Mello, and colleagues were using antisense RNA to inhibit gene expression in C. elegans. However, they decided to inject the nematodes not only with sense or antisense RNA, but also with a combination of the two. The result was a tenfold more potent silencing than the sense or antisense RNA alone. Such dsRNA-mediated gene silencing was later identified in Drosophila, zebrafish, and mammals. Achieving RNAi in mammalian cells was initially unsuccessful because dsRNA induces a powerful interferon response. This response leads to the inhibition of all gene expression and rapid cell death. Fortunately, Thomas Tuschl and colleagues discovered that introduction of synthetic short RNA duplexes did not stimulate an interferon response and instead led to sequence-specific mRNA degradation. Tuschl based the selection of the short length on the end processed forms of dsRNA in the cell, 21-22mer molecules referred to as small interfering RNAs (siRNAs). These studies provided a very powerful tool for researchers working in mammalian systems to selectively and rapidly suppress genes of interest using such synthetic siRNAs.