Zhu S W, Liu Z J, Li M, Zhu H Q, Duan L P
Department of Gastroenterology, Peking University Third Hospital, Beijing 100191, China.
Department of Biomedical Engineering, College of Engineering, Peking University, Beijing 100187, China.
Beijing Da Xue Xue Bao Yi Xue Ban. 2018 Apr 18;50(2):231-238.
To assess whether the same biological conclusion, diagnostic or curative effects regarding microbial composition of irritable bowel syndrome (IBS) patients could be reached through different bioinformatics pipelines, we used two common bioinformatics pipelines (Uparse V2.0 and Mothur V1.39.5)to analyze the same fecal microbial 16S rRNA high-throughput sequencing data.
The two pipelines were used to analyze the diversity and richness of fecal microbial 16S rRNA high-throughput sequencing data of 27 samples, including 9 healthy controls (HC group), 9 diarrhea IBS patients before (IBS group) and after Rifaximin treatment (IBS-treatment, IBSt group). Analyses such as microbial diversity, principal co-ordinates analysis (PCoA), nonmetric multidimensional scaling (NMDS) and linear discriminant analysis effect size (LEfSe) were used to find out the microbial differences among HC group vs. IBS group and IBS group vs. IBSt group.
(1) Microbial composition comparison of the 27 samples in the two pipelines showed significant variations at both family and genera levels while no significant variations at phylum level; (2) There was no significant difference in the comparison of HC vs. IBS or IBS vs. IBSt (Uparse: HC vs. IBS, F=0.98, P=0.445; IBS vs. IBSt, F=0.47,P=0.926; Mothur: HC vs.IBS, F=0.82, P=0.646; IBS vs. IBSt, F=0.37, P=0.961). The Shannon index was significantly decreased in IBSt; (3) Both workshops distinguished the significantly enriched genera between HC and IBS groups. For example, Nitrosomonas and Paraprevotella increased while Pseudoalteromonadaceae and Anaerotruncus decreased in HC group through Uparse pipeline, nevertheless Roseburia 62 increased while Butyricicoccus and Moraxellaceae decreased in HC group through Mothur pipeline.Only Uparse pipeline could pick out significant genera between IBS and IBSt, such as Pseudobutyricibrio, Clostridiaceae 1 and Clostridiumsensustricto 1.
There were taxonomic and phylogenetic diversity differences between the two pipelines, Mothur can get more taxonomic details because the count number of each taxonomic level is higher. Both pipelines could distinguish the significantly enriched genera between HC and IBS groups, but Uparse was more capable to identity the difference between IBS and IBSt groups. To increase the reproducibility and reliability and to retain the consistency among similar studies, it is very important to consider the impact on different pipelines.
为了评估对于肠易激综合征（IBS）患者微生物组成，能否通过不同的生物信息学流程得出相同的生物学结论、诊断或治疗效果，我们使用了两种常见的生物信息学流程（Uparse V2.0和Mothur V1.39.5）来分析相同的粪便微生物16S rRNA高通量测序数据。
使用这两种流程分析27个样本的粪便微生物16S rRNA高通量测序数据的多样性和丰富度，其中包括9名健康对照者（HC组）、9名腹泻型IBS患者（IBS组）以及利福昔明治疗后的患者（IBS治疗组，IBSt组）。采用微生物多样性分析、主坐标分析（PCoA）、非度量多维尺度分析（NMDS）以及线性判别分析效应大小（LEfSe）等分析方法，找出HC组与IBS组以及IBS组与IBSt组之间的微生物差异。
（1）两种流程对27个样本的微生物组成比较显示，在科和属水平存在显著差异，而在门水平无显著差异；（2）HC组与IBS组或IBS组与IBSt组的比较中无显著差异（Uparse：HC组与IBS组，F = 0.98，P = 0.445；IBS组与IBSt组，F = 0.47，P = 0.926；Mothur：HC组与IBS组，F = 0.82，P = 0.646；IBS组与IBSt组，F = 0.37，P = 0.961）。IBSt组的香农指数显著降低；（3）两种流程均能区分HC组和IBS组中显著富集的属。例如，通过Uparse流程，HC组中硝化单胞菌属和副普雷沃菌属增加，而假交替单胞菌科和厌氧短杆菌属减少；通过Mothur流程，HC组中罗斯氏菌属增加，而丁酸球菌属和莫拉菌科减少。只有Uparse流程能够找出IBS组和IBSt组之间的显著属，如假丁酸弧菌属、梭菌科1和严格梭菌属1。
两种流程在分类学和系统发育多样性上存在差异，Mothur能够获得更多分类学细节，因为每个分类水平的计数数量更高。两种流程均能区分HC组和IBS组中显著富集的属，但Uparse更有能力识别IBS组和IBSt组之间的差异。为了提高可重复性和可靠性，并保持相似研究之间的一致性，考虑不同流程的影响非常重要。
