Regional variation and diagnosis modeling of gut microbiome for inflammatory bowel diseases
10.3760/cma.j.issn.1009-9158.2018.10.007
- VernacularTitle:各地区炎症性肠病患者肠道菌群特征观察性研究
- Author:
Zhongwei WANG
1
;
Yan HE
;
Nianyi ZENG
;
Wenli TANG
;
Hongwei ZHOU
Author Information
1. 南方医科大学公共卫生学院环境卫生学系
- Keywords:
Inflammatory bowel diseases;
Gastrointestinal microbiome;
Dysbiosis;
Oxidative stress;
Forecasting
- From:
Chinese Journal of Laboratory Medicine
2018;41(10):734-741
- CountryChina
- Language:Chinese
-
Abstract:
Objective To identify microbiome biomarkers in patients with inflammatory bowel disease ( IBD) in different regions and establish predictive models , and to explore the various gut microbiota function in IBD patients .Methods The 16 srRNA gene sequences of 1510 IBD patients and 496 healthy controls were collected from China , the United States ( RISK and PRISM cohort ), Germany, India and Lithuania cohort.QIIME ( v1.9.1) was used to analyze microbiota data.LEfSe was used to identify biomarkers for IBD.Random forest method was used to establish the prediction model to distinguish IBD from HC.PICRUSt was used to predict the functional changes of gut microbiota in IBD patients .Resultsɑdiversity of gut microbial in IBD patients was significantly lower than in HC (Wilcoxon,P<0.05).The gut microbiota of IBD patients was different from HC significantly ( Adonis,P<0.05) in all of the cohort study but Indian.LEfSe analysis showed that the IBD patients from China and the U .S.cohort harbored similar dysbiosis patterns , while those from Lithuania , Germany and India have highly localized dysbiosis patterns.Generally, enterococcus was significantly increased in IBD patients in China , the U.S.and Germany cohort.Enterobacteriaceae was significantly increased in IBD patients in China and the U .S. cohort.Ruminococcus was significantly decreased in the intestines of IBD patients in China , the U.S.and India cohort.When predicting IBD status using machine learning models built on local population , the area under the curve ( AUC) was 86.48％ ±4.91％.Meanwhile, when predicting IBD status using machine learning models built on other populations , China and the U.S.had a relatively high AUC for cross-predicting, whilethe other pairs were failed when cross-applied to each other.The model established based on all samples was used to predict each population ,which showed that China , the United States ( RISK and PRISM cohort ), Germany, Lithuania and India cohorts having AUCs of 90.1％, 82.3％, 79.6％, 61.9％, 65.5％and 54.2％respectively.For functional analysis, in China, the United States (RISK and PRISM cohort ) and India cohort , glutathione metabolism and quinones biosynthesis was significantly increased in IBD patients.In China, Germany and Lithuania cohort , flagella assembly and bacterial motility proteins functions were significantly decreased in the IBD patients .Conclusions The intestinal microbiota of IBD patients from different countries could have consistent dysbiosis patterns , but geographical factors still exert a great effect on the microbiota , which needs to be further explored in subsequent studies .
