The Efficient SiRNA Screening Method Study Based On SiRNA-mRNA Binding Thermodynamic Features

RNA interference (RNAi) is gene silencing mediated by double-stranded RNA andinvolved specific enzymes. This phenomenon was discovered and named in antisense RNAinhibition experiments in the nematode worm C. elegans in1998, was later found to be awidespread biological phenomenon of gene regulation. siRNA-mediated RNA interferencemechanism can be divided into the start-up phase and the effector phase. The start-up phase isthe binding of Dicer and double-stranded RNA in the family of RNaseIII ribozyme, and cut itinto21~23nucleotides and small RNA fragments with two bases drape in3’ end, that issiRNA. Effector phase refers to the binding of siRNA and RNA-induced silencing complex(RISC) consisted several protein and helicase to single-stranded, after activating of RISC, theactivated RISC is guided by single-stranded siRNA, sequences specifically bind to the targetmRNA and cut it, lead to special decomposition of the target mRNA, and the termination ofcorresponding gene expression process.siRNA (small interfering RNA) design is an effective way to achieve RNAi, the pros andcons of siRNA sequence design will directly affect the RNAi effect, the rational design ofsiRNA sequences can improve the efficiency of gene silencing. The nucleotide sequencecontained in siRNA, the dsRNA stability of the double-stranded ends, the secondary structureof the target mRNA corresponding sequence and the binding of siRNA and RISC areassociated with the RNAi process. Due to the use of biological experiments, siRNA designrequires a lot of manpower and resources, high experimental cost, long cycle, low efficiency.Therefore, using bioinformatics and computer-aided means to optimize siRNA design hasbecome an effective way to achieve RNAi.However, not all designed siRNAs have efficient interference efficiency, the interferenceefficiency for different target position are not the same. Based on this, there are some so-calledefficient siRNA design rules, they are basically based on siRNA sequence features. In onestudy, a genome-wide siRNA is designed, using a lot of siRNA sequence features throughartificial neural network to study the inhibition rate. However, these methods only consider the sequence feather of siRNA, can not fully cover all characteristics of affecting theinterference efficiency, such as the protein binding site, cellular localization, and thesecondary structure of target mRNA, etc.There are a large number of studies have shown that the secondary structure of targetmRNA and efficiency of siRNA interference has an important relationship, especially the localsecondary structure of the binding sites has a critical role in interference efficiency. Thethermodynamics of interactions between siRNAs and target mRNA is crucial for the efficiencyof its interference. Therefore, a reliable prediction of the RNA-RNA binding energy is veryimportant. We can understand the thermodynamic energy of the RNA-RNA interactions likethis,(1) there are two kinds of energy required by opening the binding sites, they are theenergy△GM for opening the target binding site and the energy△Gs for opening siRNA,(2) the energy△Gd. Released by the binding of siRNA and mRNA. Then we can get thetotal free energy by siRNA and mRNA binding△Gt=△Gm+△Gs+△Gd.RNAup is accessibility software used to predict target resion provided by Vienna RNApackage, In this paper, using the software to add the predicted thermodynamic characteristicsto siRNA sequence features, combine with a new3-mer coding features, together withthetraditional efficient siRNA charateristics, screening efficient siRNA predicted by supportvector machine. The model can well predict efficient siRNA, the accuracy rate is up to88.12%,the sensitivity and specificity are88.35%and87.94%, respectively. Comparing withtraditional siRNA interference efficiency prediction software, this has a lot of improvement.