Phylogenetic Analysis And Non-coding RAN Prediction Based On Information Theory

With the rapid development of bioinformatics, more and morebioinformatics data has been generated. Digging out useful informationfrom the bioinformatics data can improve the study of life sciences,provide a theoretical support for the treatment of diseases and drug design.In this paper, we analyze the distribution of nucleotides and build theInter-Nucleotide Distance model to distinguish different genes. The maincontents are followed:1. Proposed a new method for phylogenetic analysis. We proposed aimproved Inter-Nucleotide Distance model. This model calculates thedistribution of the other three nucleotides between the same nucleotide,which is more accurate to analyze short DNA sequences (less than100000bp). Associated with log-correlation distance and Euclideandistance, we analysis the64vertebrate mitochondrial genomes and70polyomavirus genomes and construct phylogeny tree, the results coincidewith those obtained using traditional methods. Hence the phylogeneticmodels or methods proposed by us are reliable and stable. They aresignificant for the phylogenetic analysis.2. Proposed a new method for non-coding RNA prediction. Non-coding RNA prediction usually use the Statistical characteristics of theRNA sequences.However, the Statistical characteristics can not fullyexpress the characteristics of RNA sequence. Our Inter-NucleotideDistance model can convert the non-coding RNA sequence into anumerical sequences and use for non-coding RNA prediction. Associatedwith support vector machine, we use a set of E-coli-K12non-coding RNAsequences in prediction analysis. Compared with other methods, ourmethod calculate more sequence information and have a better result,which is feasible in the prediction of non-coding RNA sequences inprokaryotes.