Research On Fast Alignment-free Methods For Microbe Classification And Evolution

Microbes are the most widely distributed organisms in nature with simple structures in which viruses and bacteria are the most common microorganisms.Microbes are the main participants in the material cycle of nature,and can have a huge effect on human activities.With the rapid development of sequencing technology,many kinds of microbial genomes have been sequenced.Using genome comparison methods,we may classify these organisms and explore their origin and evolution.It is a crucial step for understanding of microbial functions,prevention and diagnosis of infectious diseases caused by microorganisms.Traditionally,multiple sequence alignments are very effective to construct accurate evolutionary relationships.However,due to high mutation rates of viral sequences,insertion and deletion of genes and gene horizontal transfer,the alignment methods may lead to inaccurate results.In addition,the computation time for multiple sequence alignment methods is proportional to the length of a sequence,which make theses methods unsuitable for classification of bacteria and viruses with large genomes.In this paper,based on proteomes of viruses,we used the 60 dimensional vector method to make classification for more than 4000 viruses in seven Baltimore classes and construct their evolutionary relationships.The results show that our method can achieve very high accuracy.We also compared the classification by proteomes with that by genomes.We find that classification by proteome data can achieve better results when reliable protein sequences are sufficient,otherwise the results are also comparable.Using the natural vector method,we also built phylogeny of the emergent Zika virus and other flaviviruses.The conclusion shows that Zika virus is originated from Africa,then spreads to Asia,the Pacific and throughout the Americas.Finally,we propose a new feature vector method for constructing the evolutionary relationships of species.Our vector method not only utilizes the distribution of nucleotides in sequences,but also includes the biochemical properties of sequences.The conclusion tested on several microbial data sets shows that our method is very fast and can be used to accurately establish evolution of organisms.