RESUMO
Short interfering RNA (siRNA) has been widely used for studying gene functions in mammalian cells but varies markedly in its gene-silencing efficacy in mammalian genes. The recently reported guidelines for selecting effective siRNA target sequences are not always useful for selecting highly effective siRNA sequences for many other mammalian genes because there are only a few consistencies among them. Hypothesizing that the positional nucleotide occurrence trends play an important role in effective gene-silencing, we examined 361 effective siRNA sequences from 227 different mammalian cDNAs in the literature and found got several nucleotide features different from the ones used in the previous guidelines. Here we first explain the problems concerning the previous guidelines from the qualitative and quantitative points of view. Then after clarifying the requirements for effective siRNA designs, we describe a new method based on a gene degradation measure defined by positional features of specific significant nucleotides. Testing the method on human cyclin B1 confirmed that it selected highly effective gene-silencing sequences and also indicated that it would be useful for other genes. It will therefore be useful for selecting new siRNA target sequences for mammalian genes.
