Improving model predictions for RNA interference activities that use support vector machine regression by combining and filtering features.

ABSTRACT:

Background

RNA interference (RNAi) is a naturally occurring phenomenon that results in the suppression of a target RNA sequence utilizing a variety of possible methods and pathways. To dissect the factors that result in effective siRNA sequences a regression kernel Support Vector Machine (SVM) approach was used to quantitatively model RNA interference activities.

Results

Eight overall feature mapping methods were compared in their abilities to build SVM regression models that predict published siRNA activities. The primary factors in predictive SVM models are position specific nucleotide compositions. The secondary factors are position independent sequence motifs (N-grams) and guide strand to passenger strand sequence thermodynamics. Finally, the factors that are least contr

SUBMITTER: Peek AS 

PROVIDER: S-EPMC1906837 | biostudies-literature | 2007 Jun

REPOSITORIES: biostudies-literature