Microbiome Big Data Evolution-Ecological Fusion Visualization Platform Construction And Algorithm Research

The footprints of microbes are everywhere in the world.From oceans to volca-noes,from high mountains to indoors,even the number of microbes in the human body is greater than the number of cells in the human body.It can be said that microbes are in-volved in most of the life processes on earth.Understanding the microbiome community play a vital role in the development of earth sciences,life sciences,biomedicine,agri-culture,bioenergy and other industries.With the rapid development of next-generation gene sequencing technology,the cost of sequencing has decreased year by year.Scien-tific researcher can more easily access the data on alkaloid sequences,and research on more and more related domain.The development of metagenomics is so fast that we need to process mounts of data.Traditional biological experiments are difficult to meet the needs of metagenomic research,so we need to use computers to complete subse-quent data analysis.Due to metagenomic data is very complex,intuitively we humans can't observe any information,so we need a good enough visualization platform to pro-cess metagenomic data and visualize it for us humans to read.We have completed the construction of the microbiome big data evolution-ecological fusion visualization plat-form for microbiological data analysis.There are three main parts here:a web-based metagenome visualization platform that can interact with users,research on supervised machine learning technology based on optimal segmentation of evolution trees,and re-search on network structure learning technology of microbial species ecological func-tion.The basic function of the metagenomics visualization platform is to visualize the microbiome data,including heat maps,phylogenetic trees,species community composi-tion,enrichment analysis,and microbial community diversity visualization.Compared with other metagenome visualization platforms,our part is characterized by our tool that can filter the branches of certain evolutionary trees to achieve the simultaneous change of heat maps,evolutionary trees,and species community composition maps to only represent the currently selected.The branch information is more convenient for us to observe the branch information.Research on supervised machine learning technology based on evolutionary tree optimal segmentation and research on microbial ecological function network structure learning technology are two methods for extracting features from microbial data for subsequent data analysis.We intergrate this method to our plat-form.Research on supervised machine learning technology based on evolutionary tree optimal segmentation is a feature of our platform.Traditional metagenomics data anal-ysis is based on OTU.We introduced the evolutionary tree information and used the idea of greedy search.Obtained a combination of some OTU ancestor nodes and other OTUs,and used them as the objects of our subsequent data analysis.On the data set we tested,its results are more excellent than the traditional data analysis using only OTU.Similarly,the study of microbial species ecological function network structure learning technology is another feature of our platform.The similarity matrix is obtained through the similarity relationship between OTUs.This can be regarded as a matrix representa-tion of a graph,and then using graph embedding technology Generated a second-order similarity matrix,Then use spectral clustering to cluster into several classes,we can use several classes as a module,each module contains a number of OTU,so the data of each module can be regarded as a feature,we These new features can be used for subsequent data analysis.Finally,our current platform can apply research on human chronic dis-ease data.Users upload data to our platform,and our platform can quickly visualize the data.Users will obtain statistically what OTUs are obviously associated with diseases.Through the use of our phylogenetic tree-based supervised machine learning technology and microbial ecological function network structure learning technology,users will see which OTUs combined have a significant auxiliary role in the diagnosis of diseases.