Tuschl's algorithm provides a potential list of siRNAs, and in order to narrow down this list, other algorithms have been developed. These algorithms further constrain siRNA design by applying additional rational design principles in the hope of rendoring the resultant siRNAs more efficient and specific.
|
An Algorithm for selection of functional siRNA sequences M. Amarzguioui, H. Prydz. Biochem. Biophys. Res. Commun. 316 (4) (2004) 1050-1058. This algorithm used structural/sequence properties of siRNA to determine the properties that determine high efficacy levels. Predictors of functional siRNA are as follows: high differential between A/U content in 3’ end and 5’ end, the inclusion of S1, A6, W19, the exclusion of U1, G19. The authors also recommend a GC content of around 32-53% and claim that mRNA secondary structure is not important. |
Guidelines for the selection of highly effective siRNA sequences for mammalian and chick RNA interference K. Ui-Tei, Y. Naito, F. Takahashi, T. Haraguchi, H. Ohki-Hamazaki, A. Juni, R. Ueda, K. Saigo. Nucleic Acids Res. 32 (3) (2004) 936-948. The results of this paper focused on sequence properties of siRNA which yielded high efficacy levels. The findings are that efficient siRNAs have the following properties: 5’ antisense strand end should be A or U, 5’ sense strand end should be G or C, 4 out of 7 nucleotides in the 5’ terminal antisense strands are A or U, and no GC stretch of greater than 9nt should exist. The authors also found that the “target sequence is much more important for target recognition by siRNAs than the secondary structure”. |
A library of siRNA duplexes targeting the phosphoinsitide 3-kinase pathway: determinants of gene silencing for use in cell-based screens A. Hsieh, R. Bo, J. Manola, F. Vazquez, O. Bare, A. Khvorova, S. Scaringe, W. Sellers. Nucleic Acids Res. 32 (3) (2004) 893-901. This paper also examined sequence features to classify effective RNAs. The authors found that targeting the middle of a coding sequence was ineffective. In addition targeting 3’ UTR is comparable to targeting a coding sequence. Pooling duplexes were found to be efficient in knocking down genes. There were strong nucleotide preferences at 11 (G or C) and 19 (T). |
Rational siRNA design for RNA interference A. Reynolds, D. Leake, Q. Boese, S. Scaringe, W.S. Marshall, A. Khvorova. Nat. Biotechnol. 22(3)(2004) 326-330. This algorithm uses sequence features to determine RNA efficiency. There are 8 characteristics for siRNA functionality: low G/C content, a bias for low internal stability at the sense strand 3’-terminus, lack of inverted repeats, sense strand base preferences (positions 3, 10, 13, 19). An algorithm which incorporates all of these properties improves gene knockdown. |
Asymmetry in the assembly of the RNAi enzyme complex D.S. Schwarz, G. Hutvagner, T. Du, Z. Xu, N. Aronin, P.D. Zamore. Cell 115 (2003) 199-208. This paper looks at the assembly of siRNA into the RISC complex. It shows that 2 strands are not equally likely to be incorporated. Instead, which strand is incorporated depends on “absolute and relative stabilities of base pairs at 5’ ends of 2 siRNA strands” Asymmetry is important for siRNA function. |
Functional siRNAs and miRNAs exhibit strand bias A. Khvorova, A. Reynolds, S.D. Jayasena. Cell 115 (2003) 209-216. It was found that functional duplexes display a lower internal stability at 5’-AS end than non functional duplexes. Their work suggests that thermodynamic properties of siRNA “play a critical role in determining molecule’s function and longevity, possibly biasing the steps involved in duplex unwinding and strand retention by RISC”. |
Improved and automated prediction of effective siRNA A. Chalk, C. Wahlestedt, E. Sonnhammer. Biochem. Biophys. Res. Commun. 319(1)(2004)264-274. The authors tried to find properties of effective siRNAs. The seven properties that siRNAs were rated on are a total hairpin energy <1, a middle binding energy < 13, binding energy difference (for antisense 5’ end – sense 5’end) is <0, energy difference between -1 and 0, GC content between 36-53%, antisense 5’ end binding energy <9, sense 5’ end binding energy in range 5-9. Each siRNA is given one point for each criteria it satisfies so the scores range from 0-7. A cutoff score of 6 yields 52% high efficacy which is over two fold increase from the whole sample. They developed software which allows the user to design siRNA based on different people’s results and algorithms |
The gene-silencing efficiency of siRNA is strongly dependent on the local structure of mRNA at the target region K. Luo, D. Chang, Biochem. Biopys. Res. Commun. 318 (1)(2004)303-310. This study focuses on selecting on effective target region for the siRNA. The explored property is the secondary structure of the target site. The H-b index is defined which measure the average number of hydrogen bonds formed in a region. The authors find that a lower H-b index results in higher RNAi efficiency which they predict is due to the siRNA having easier access to unfolded mRNA. An advantage to this idea is that the correct secondary structure of mRNA does not need to be determined. Instead an average H-b index is determined for all predicted secondary structures. The authors also find that hairpin loops in the target region decrease efficiency since it will be harder for siRNA to bind in the area. They also possibly recommend against picking target regions at the 5’ or 3’ ends of mRNA. |
| next |