C.elegans coding and non-coding zones classification using multifractal features and support vector machine

Abstract

Research on exons and introns is important to the structure knowledge, regulation, and genes’ evolution. The multifractal analysis allows the genetic structure to be studied on varying scales. This paper presents a novel approach to characterize and classify C.elegans exons and introns sequence based on multifractal features and the Support Vector Machine (SVM) classifier. First, we transform the genomic sequence into an FCGR image, and then the multifractality spectrum is calculated by applying the box-counting method. Finally, we use the SVM classifier by testing different kernels. Our classification method has shown promising results for the learning and testing rates exceeding 75%