Maximum sequence alignment fails to predict off-targeted gene regulation by RNAi

We have employed various sequence alignment algorithms and scoring techniques to determine whether current computational tools accurately predict genes that will be off-targeted by the RNA interference (RNAi) pathway. Our studies show that distributions of maximum alignment scores for off-targeted and untargeted genes are statistically indistinguishable, indicating that maximum complementarity by itself is an unsatisfactory predictor of off-targeting. Interestingly, a highly significant association was observed between off-targeting and exact complementarity between the seed region (bases 2-7) of siRNA and their off-targeted genes. This pattern has been previously recognized in microRNA-mediated gene knockdown and suggests a distinctive role for the 5 terminus of these strands in RNAi-triggered gene suppression.