Network Construction Analysis Of Gut Microbiota In Patients With T2DM

With the development of biomedicine level.Combining medical information with computer science.The interdisciplinary research is more and more extensive.In the field of bioinformatics,the study of intestinal flora has gradually come into public view.With the updating of sequencing technology,a large number of high-throughput intestinal flora data have emerged to participate in the study of diseases.This requires us to analyze the important gut bacteria in the vast amount of data.These key strains may provide new ideas for treating diseases.At present,more emphasis is placed on the statistical significance analysis of diversity indicators in the study of the relationship between intestinal flora and diseases.This method can not effectively show the significant differences between normal and sick people's microbiome,but can only show that the disease and intestinal microflora are related.In order to reveal the differences of intestinal flora among different populations,network analysis was used to construct the network and identify the key nodes in the network.These key nodes represent key species in the gut.Most studies use arbitrary thresholds when constructing networks.The microbial interaction network obtained at this time is more subjective and may not reflect the actual situation.And there is a lack of verification of the validity of the network after the construction of the network.The main idea of network analysis is to calculate the correlation coefficient between each node in the network.The microbial interaction network was constructed according to the coefficient values in a certain range.Based on the above basis,two experiments were conducted in this paper to analyze the differences in intestinal flora between type 2 diabetes patients and normal population,mainly including:In the first part,the network is constructed by the custom threshold.The correlation coefficient matrix and p-value matrix between OTU are calculated by Pearson correlation coefficient method in the WGCNA package of R language,and then the pvalue is corrected by FDR method.After that,the correlation coefficient R is 0.66,0.88 and 0.99 respectively in the case of P<0.05 to construct the network.After a power-law distribution of the network node degree of fitting and random network authentication network characteristics,found people to determine the threshold is 0.88 build group,type 2 diabetes type 2 diabetic diarrhea group and the control group's network in the network size,the power-law distribution index R2,network coherence characteristics(clustering coefficient and average path length)than other two threshold of building the network effect is better.The second part,using the MENA molecular ecological network analysis of RMT(random matrix theory)threshold detection algorithm to calculate threshold and descending order,choose the poisson distribution and poisson distribution chi-square test P value is greater than the threshold value of 0.05 0.66 build a network of three kinds of crowd,after evaluation of the effectiveness of the network,and the custom under the threshold of the network topology characteristics were analyzed,the results found by threshold detection algorithm to determine the threshold of constructing network in network scale,R2 value,network coherence characteristics were higher than the custom threshold of building network,network effect is better.Secondly,in the process of constructing the intestinal flora network,this article found that no matter which method is constructed,the flora interaction network will lose a large number of nodes and edges.These nodes and edges do not meet the requirements of network construction and are therefore discarded.However,the interaction mechanism of the intestinal flora is still unclear,and network analysis is a mathematical inference method,so the flora data cannot be fully displayed.It is necessary to understand the size of the intestinal flora data after the network is constructed,and how much nodes and edges are lost in the final network.This paper proposes a new network attribute,the edge loss rate(S /L,Saving/Lost),S represents the number of edges in the network,L represents the number of edges between pairs of nodes that are lost after a certain threshold is selected to construct the network,and the edge loss rate is expressed in networks constructed under different thresholds stable.After determining the optimal threshold to construct the network,the MCODE method in Cytoscape software was used to identify the core intestinal microbiome in the overall network,and then the core nodes were determined according to the connectivity Zi value in the network node attribute module in descending order to find the key species.The analysis found that the intestinal microorganisms of the three types of people contained a large number of similar bacteria,including Firmicutes,Bacteroidetes,Actinomyces and Proteobacteria.However,the intestinal flora of patients with type 2 diabetes mellitus complicated by diarrhea was more complex.Botulinum bacteria,Pasteuria,Richenia,and Violetomonas were the key bacteria in the intestinal flora of patients with type 2 diabetes mellitus complicated by diarrhea,while these bacteria were not included in the core intestinal microbes of the other two groups.Finally,the correlation between intestinal flora and liver and kidney biochemical indexes in patients with type 2 diabetes mellitus was statistically analyzed.The results showed that the increase of serum bilirubin in patients with type 2 diabetes mellitus complicated with diarrhea was correlated with the decrease of intestinal flora species diversity.The contents of Bacteroides,Trichoderma and Prevotella in patients with type2 diabetes mellitus complicated with diarrhea decreased,but the contents of rumen bacteria increased significantly.In addition,there was a positive correlation between pilorucella and rumen bacteria and serum bilirubin,which was related to the decrease of TG,but not to the increase of BUN.There was significant correlation between liver and kidney indexes and intestinal flora in patients with type 2 diabetes mellitus complicated with diarrhea.