Using Random Perturbation Method to Improve Efficacy Prediction of siRNA Sequences

Short interfering RNAs (siRNAs) can be used to silence gene expression. The potential application of this research opens the door for the development of the future generations of effective drugs. The question of how to design effective siRNAs is still under intense study. Our motivation is to improve the accuracy of the simple linear model studied in [1]. We propose a new random perturbation method that predicts most active siRNA candidates with high accuracy. The contribution of this paper is two-fold. First, we developed the random perturbation method which improved the prediction of siRNA sequence efficacy in [1]. Our method is easy to implement and runs efficiently. Second, we developed a new method; we named it alpha method that can linearly optimize different contributions from different weight sets.