A Sequence-based Predictor for Identifying DNase Hypersensitive Sites Via Physical-chemical Property Matrix

DNase I hypersensitive sites(DHSs) are important signs of regulatory DNA, moreover by most accounts have supported the research findings of all sorts of cis-regulatory elements containing silencers, promoters, locus control regions, enhancers and insulators. Therefore, DHSs plays a crucial role for deciphering the function of noncoding genomic regions. At present, genome sequences have been growing fast, it is very crucial to develop new-style computational means for effctively and rapidly distinguishing DHSs. In this study, we development a new predictor, in which the DNA sequence sample were described by a very novel mode obtained from a physicochemical property matrix (PCM) by means of an array of cross-covariance and auto-covariance trans. It was demonstrated via rigorous jackknife test that compared with the existing predictor for the same purpose, our predictor achieved remarkably higher rates, showing that our predictor can become a more ideal tool for distinguishing DHSs.