Repeating experiments is crucial for data evaluation. However, one does not expect having to re-evaluate all the genome-wide RNA interference (RNAi) screens that were carried out in Drosophila melanogaster cell cultures.

The groups of Philip Beachy (Johns Hopkins University School of Medicine, USA) and Norbert Perrimon (Harvard Medical School, USA) did exactly that. Unexpectedly, they found that off-target effects (OTEs) can occur in RNAi experiments using long double-stranded RNA (dsRNA) sequences.

OTEs have been recognized as a problem for RNAi screens that are carried out using short RNAs. However, genome-wide D. melanogaster libraries contain dsRNAs that consist of hundreds of bases, which were not thought to produce OTEs. “Initially, no one in the field ... had seen an off-target effect, but as more screens were done, we realized these things could happen”, said Perrimon (The Scientist, 11 Sep 2006).

Suspicion arose when Yong Ma, a postdoctoral researcher in the Beachy laboratory, attempted to validate seven promising candidates for novel components of the Wingless pathway that were identified using a library of more than 21,000 dsRNAs in D. melanogaster cells. Ma found that all candidates contained short regions of homology with armadillo, a well known gene of this pathway. Looking back at another published RNAi screen on the Wingless pathway, Ma realized that ∼60% of the negative regulators that were identified shared a repeated trinucleotide sequence, CAN, that is found in only 5% of the library.

Perrimon found evidence of OTEs in a retrospective analysis of 30 RNAi screens. The false positive rate is variable. Whether false or suspect, according to Perrimon “... these need to be repeated with new dsRNAs, which we are doing now” (The Scientist, 11 Sep 2006).