| Liver cirrhosis a chronic liver disease.There is marginal difference between patients with liver cirrhosis at early stage and healthy cohort about liver function issue.As liver disease syndrome occurs,such as varices,hepatic encephalopathy,ascites,patients is diagnosed as decompensated liver cirrhosis finally end up with end stage liver disease,i.e.ESLD.It is well recognized that liver disease is associated with alternations of gut microbiota.Translocation and disturbance of gut microbiota will simulate inflammatory factors then induce complications such as hepatic encephalopathy and sepsis,also,gastric acid reduction leads to an alteration of species richness and diversity in lower intestinal microbiota.Our previous study showed significant increase of bacteria from Enterobacteriaceae and Veillonellacea in patients with liver cirrhosis(LC),however,Lachnospiraceae decreased significantly,most of pathogens enriched in liver cirrhosis origin from upper gastrointestinal tract.Besides,genes involved in ammonia synthesis is enriched in LC cohort,which might be associated with blood ammonia level then burden liver functions.To further explore role of microbiota in liver cirrhosis process,we use whole genome shotgun sequencing data of stools to investigate its association with disease.MethodsWe enrolled healthy controls(H),patients with decompensated liver cirrhosis(decLC)and compensated liver cirrhosis(cLC).With taxonomy profile and orthology composition generated by WGS data,to parse factors might be associated with disease process(H>cLC>decLC).Also,patients with LC were divided into Death cohort and Alive cohort with a follow-up study,we also parse factors might be associated with death.Finally,about data analysis method part,we update algorithm in taxonomy composition parse and function annotationResults and discussionMicrobiota diversity significantly decrease in deCLC cohort,compared with CLC and H.Also,microbiota composition of CLC and H is more similar.A total of 33 species and 63 orthologies,positively associated with disease process,10 species and 123 orghologies negatively associated with disease process.These different species could be clustered into 2 microbial networks.About the function pathway,Biosynthesis of amino acids,Methane metabolism and Aminoacyl-tRNA biosynthesis negatively associated with disease process while Nitrogen metabolism positively associated with disease process.Distribution of some well-recognized probiotics,including Akkermanisa muciniphila and Bifidobacterium longum,showed no difference in CLC and H.In a nutshell,microbiota is disturbed in CLC and the homeostasis imbalance is associated with disease process.Composition of microbiota showed no significant difference in Alive and Death,however,microbial diversity increased in Alive cohort.Some important probiotics and Butyrate-producing bacterium such as Bifidobacterium longum,Oscillibacter valericigenes and Ruminococcus obeum,decreased in Death cohort.Also,these three species is also associated with disease process.We use 43 species associated with liver cirrhosis process and random forest(RF)model to construct a LC diagnosis model and calculate probability of disease(Probability of disease,POD)for each sample.POD is also significantly associated with disease process.An independent validation showed this model have 0.7925 precision rate to distinguish CLC and H,0.8679 precision rate to distinguish LC and H.POD showed some potential to distinguish CLC and deCLC(auc is 0.6474),however,it do not work for Alive and Death classification.We develop an algorithm,multiple best local alignments for function annotation of genes,a simulation test of COG database showed the precision rate could be 0.9215,compared with 0.2561 by using routine annotation method.Also,a uniCo(Unique Coefficient)algorithm was developed to construct taxonomy profile and gene profile,a simulation test of mock community showed the correlation could be increased from 0.57 to 0.809(Pearson correlation test). |
PDF Full Text Request