Classification DNA sequences using Gap–Weighted subsequences kernel

Abstract

The aim of this paper is to show experimental results of classification DNA sequences using gap–weighted subsequences kernel including the assess the expected error rate of a classification algorithm. The process involve a type of kernel specific with a classification algorithm for learn to recognize sites that regulate transcription, sites that can be detected in the laboratory as DNaseI hypersensitive sites (HSs) on DNA sequences. The classification algorithm is support vector machine (SVM), which learns by example to discriminate between two given classes of data. The DNA sequences are converted using gap–weighted subsequences kernel in a matrix kernel, which is processed by the classification algorithm to produce a model with the which we can predict the classification of new examples. It is important to know that a high accuracy with computational methods for the identification of the DNaseI hypersensitive sites would to help to speed up the functional annotation of the human genome