Research On SiRNA Target Screening Method On MRNA

Using double-stranded siRNA to knock down target mRNA is an effective way to achieveRNA interference.The length of mRNA is much longer than that of siRNA,so one mRNA can be paired with multiple siRNAs.SiRNAs targeting different positions of the same mRNA have very different efficiencies in knocking down their targets,and the more efficient siRNAs account for less.It is the target site that determines the efficiency of siRNA.Therefore,how to screen an mRNA for a site that is easy to bind to siRNA is the most concerned issue inRNAi research.This article first predicts the efficiency of siRNA knocking down the target at each site on the mRNA,and then considering both the predictive value of knockdown efficiency and the possibility of off-target at the same time,the site with high predictive value and low off-target risk is selected as the final Target.There are currently a variety of siRNA target screening methods.Most of these methods use machine learning algorithms to predict siRNA knockdown efficiency to screen targets.However,their feature extraction still rests on the siRNA and mRNA sequence itself and some thermodynamic parameters,and the accuracy of prediction needs to be improved.This article has conducted a comprehensive analysis of the existing siRNA experimental data,and in-depth studied various factors that affect the efficiency of siRNA knockdown.On the basis of the existing research,the secondary structure of mRNA and the secondary structure of siRNA are further explored,and the partition function algorithm for predicting the secondary structure ofRNA is used to calculate the pairing probability of each base in the single-strandedRNA,and the secondary structure ofRNA is realized.Description of structural features.Then combine the existing features with the new features proposed in this article to filter out effective features,and the Light GBM model was established to predict siRNA targets.Compared with the existing prediction methods,it has obtained better prediction results on Huesken,Vickers,Reynolds and other data sets.In addition,a knockdown experiment was performed to test the prediction effect of the model for the PDCD1 gene.The experimental results further show that the target predicted in this paper has achieved better knock-down efficiency.On the basis of the researched prediction method,this article made an siRNA target screening website for specific mRNA sequence: http://sbo-bio.com.cn/2007/.After inputting the mRNA sequence,the website can calculate a suitable siRNA target through the background program.