Feature selection for siRNA efficacy prediction using natural computation

RNAi is a naturally occurring, highly conserved phenomenon of RNA mediated gene silencing among the multicellular organisms. Currently, RNAi has been successfully applied in functional genomics, therapeutics and new drug target identification in mammals and other eukaryotes. The uniqueness lies in sequence specific gene knock down which made RNAi an indispensible technology. In the mechanism of RNAi, small 21-25 bp long ,double stranded endogenously or exogenously generated RNA molecule via DICER; a RNAase III enzyme, resulting in the formation of multidomain RISC structure. The slicer activity of RISC cleaves the complementary mRNA and blocks the transcription process, consequently the gene function. Many siRNA design tools have been developed following the general guidelines like rational design criteria with their uniquely defined features. These guidelines are based upon sequence to thermodynamic features of siRNA which plays crucial role in the effective designing of siRNA. But no tool provides the accurate gene specific siRNA sequences with absolute efficacy. Furthermore, selection of suitable siRNA from the bunch of predicted sequences is the most intriguing challenge even now. So identification of features for effective siRNA design is the very smoldering issue to be solved in the present scenario since feature selection is computationally a NP hard problem. In this paper, we have presented an ant colony optimization based meta-heuristic methodology to identify the features of siRNA up to the considerable amount of accuracy and further the results are analyzed using linear regression and ANCOVA methods. It has been observed that sequence based features are comparatively significant than thermodynamic features. Further the analysis reveals that both sequence and thermodynamic features are simultaneously important in the effective designing of siRNA.