Detection Of Horizontally Transferred Genes By Using Biological Function Category

Horizontal gene transfer(HGT) which also known as lateral gene transfer(LGT) refers to the DNA fragment exchange of different individual organisms or organelle(within the same cell). Horizontal gene transfer event is an important driving force to promote the diversity of microbial genome evolution, showing a profound impact.With the increasing of the amount of available genomic data, the importance of horizontal gene transfer to the diversification of organism genomes has been widely recognized, which makes the prediction methods of horizontally transferred genes to be more significant. While the earliest methods were focused on simply establishing horizontally gene transfer as a major influence in genome evolution, the increased sophistication of next-generation methods has provided greater confidence that the individual genes suspected as having been introduced by gene transfer are actually of foreign origin. Abnormal gene for gene prediction method of horizontal transfer comes gradually into two categories: Phylogenetic methods and parametric methods. Phylogenetic methods examine the distribution of genes among genomes to detect those with apparently anomalous patter ns. In contrast, parametric methods are also known to be composition-based approaches which examine the proper ties of genes within a single genome to find compositionally anomalous genes. Phylogenetic methods is time consuming and not sufficiently robust. Parametric methods can be in a single genome, do not require a priori knowledge of a large number of sequences, such as pairing orthologous, small amount of calculation and has much room for improvement. Therefore, Parameter methods deserve our study and this paper is based on it.However, there exist two problems in the conventional feature methods(1) the predictions of all those methods are based on the whole genome, which may inevitably result in the inaccurate expression of the entire genome by a single feature and therefore lower the prediction accuracy. So in this paper, in order to achieve more accurate expression of the genome and improve prediction accuracy, we divide the genome into several gene function categories, and select different features according to different function categories;(2) only a few genomes are used commonly for the validation of those genomic prediction methods. To avoid the short of enough data leading to the lack of the university of the experiment method,, we selected all the 326 genomes that we find for our experiment. The results show that horizontal transfer genes have a certain preference for their functions. And a corresponding optimal feature can be found for a same biological function category for effective expression. Accordingly, we provide our recommended feature expression for each function category, which has a strong guiding significance for the prediction of horizontal transfer genes in prokaryote genomes.In this paper,our methods are the first time to predict transfer gene from biological functional genes of the genomes. Especially, each function class recommendation feature expression is given in the last, which is necessary and significant for the prediction horizontal transfer gene in the amount of data of the genome.