High-throughput Sequencing Data Mining For Human Skin Microbial Communities

Human skin is the first line for defending against pathogens and meanwhile the skin surface is symbiotic with a large number of micro-organisms in this diverse environment,including bacteria,fungi and viruses.If taken skin as an ecosystem,it is an overall integration of all kinds of micro-organisms including skin epidermis tissue cells and a variety of secretions,maintaining delicate balance between microbial communities and skin tissue environment,breaking of which will incur skin infections or disease.Therefore,it's of great importance to study skin microbial community and get to know relations between different skin microbial communities to figure out difference of various microbial communities.Based on high throughput sequencing data samples of different microbial communities of human skin.First adopted interpolation Markov model for background sequencing modeling to alignment-free method of short k-tuple,and relations among different communities are identified through unsupervised clustering analysis.Then,based on the class information obtained by clustering,the long k-tuple(k>30bp)sequence was used as the features to supervise the classification,and the specific tuples in different metagenomic samples are identified.Finally,specific sequences were registered to related genome database through BLAST to figure out relative differential micro-organism and biological significance of micro-organism community features were identified.This paper analyzed skin micro-organism communities based on unsupervised clustering method and designed seven experiments,carrying out deepened exploration on relations of singular individual and multi-individual different micro-organisms.These figured out that difference of micro-organism communities of different parts on singular individual is most obvious and difference on left and right side is greater than that of different skin types and most samples can keep relative stability in certain time instinctively.Either in singular individual or multi-individual,sebaceous skin and moist skin can be well distinguished and difference among hybrid individual is greater than that among different individuals.Based on skin grouping information of different communities through unsupervised clustering,the aforesaid classification and characteristics extract method is applied to human skin micro-organism community sample data to distinguish skin samples on different parts of left and right side and identify differential long sequencing characteristics of sebaceous skin and moist skin.Analysis of identified difference characteristics figures out that great difference exist on different parts of left and right side and different type of skin micro-organism composition.Large amount of pseudomonas exist on skin sample of left parts while much more propionibacterium acnes exist on skin sample of right sides.Species variety is of higher level for skin micro-organism on left and right side,but of lower level on genus.In experiments distinguishing sebaceous skin and moist skin samples,much more fungus such as Malassezia and Sporisorium exist on surface of sebaceous skin while large amount of Ralstonia,Haemophilus,Escherichia and Sulfolobus in Archaea community exist on surface of moist skin.We also found out that variety of micro-organism of sebaceous skin is much lower on phylum while moist skin is on the contrary.All these indicate that different body parts and ecological environment will facilitate growth of different micro-organism and then present obvious difference.These detailed biological significance are of important reference significance to exploring micro-organism composition of human skin and related disease genes.