Division of Urogynecology & Reconstructive Pelvic Surgery, Division of Reproductive Sciences, Department of Obstetrics & Gynecology, Duke University Medical Center, Durham, NC, United States.
Department of Statistical Science, Duke University, Durham, NC, United States.
Front Cell Infect Microbiol. 2022 Jul 8;12:789439. doi: 10.3389/fcimb.2022.789439. eCollection 2022.
An approach for assessing the urinary microbiome is 16S rRNA gene sequencing, where analysis methods are rapidly evolving. This re-analysis of an existing dataset aimed to determine whether updated bioinformatic and statistical techniques affect clinical inferences.
A prior study compared the urinary microbiome in 123 women with mixed urinary incontinence (MUI) and 84 controls. We obtained unprocessed sequencing data from multiple variable regions, processed operational taxonomic unit (OTU) tables from the original analysis, and de-identified clinical data. We re-processed sequencing data with DADA2 to generate amplicon sequence variant (ASV) tables. Taxa from ASV tables were compared to the original OTU tables; taxa from different variable regions after updated processing were also compared. Bayesian graphical compositional regression (BGCR) was used to test for associations between microbial compositions and clinical phenotypes (e.g., MUI versus control) while adjusting for clinical covariates. Several techniques were used to cluster samples into microbial communities. Multivariable regression was used to test for associations between microbial communities and MUI, again while adjusting for potentially confounding variables.
Of taxa identified through updated bioinformatic processing, only 40% were identified originally, though taxa identified through both methods represented >99% of the sequencing data in terms of relative abundance. Different 16S rRNA gene regions resulted in different recovered taxa. With BGCR analysis, there was a low (33.7%) probability of an association between overall microbial compositions and clinical phenotype. However, when microbial data are clustered into bacterial communities, we confirmed that bacterial communities are associated with MUI. Contrary to the originally published analysis, we did not identify different associations by age group, which may be due to the incorporation of different covariates in statistical models.
Updated bioinformatic processing techniques recover different taxa compared to earlier techniques, though most of these differences exist in low abundance taxa that occupy a small proportion of the overall microbiome. While overall microbial compositions are not associated with MUI, we confirmed associations between certain communities of bacteria and MUI. Incorporation of several covariates that are associated with the urinary microbiome improved inferences when assessing for associations between bacterial communities and MUI in multivariable models.
评估尿微生物组的一种方法是 16S rRNA 基因测序，其中分析方法正在迅速发展。对现有数据集的重新分析旨在确定更新的生物信息学和统计技术是否会影响临床推论。
先前的一项研究比较了 123 名混合性尿失禁（MUI）女性和 84 名对照者的尿微生物组。我们从多个可变区获得未处理的测序数据，从原始分析中处理操作分类单元（OTU）表，并对临床数据进行去识别处理。我们使用 DADA2 重新处理测序数据以生成扩增子序列变异（ASV）表。ASV 表中的分类单元与原始 OTU 表进行比较；更新处理后的不同可变区的分类单元也进行了比较。贝叶斯图形组成回归（BGCR）用于测试微生物组成与临床表型（例如 MUI 与对照）之间的关联，同时调整临床协变量。使用几种技术将样本聚类为微生物群落。多变量回归用于测试微生物群落与 MUI 之间的关联，同时再次调整潜在混杂变量。
通过更新的生物信息学处理鉴定的分类单元中，只有 40%最初被鉴定出来，尽管通过两种方法鉴定的分类单元在相对丰度方面代表了测序数据的>99%。不同的 16S rRNA 基因区域导致不同的恢复分类单元。通过 BGCR 分析，微生物组成与临床表型之间存在低（33.7%）关联的可能性。然而，当微生物数据聚类为细菌群落时，我们证实细菌群落与 MUI 相关。与最初发表的分析相反，我们没有按年龄组识别出不同的关联，这可能是由于在统计模型中纳入了不同的协变量。
与早期技术相比，更新的生物信息学处理技术可恢复不同的分类单元，尽管这些差异大多存在于占据微生物组小部分的低丰度分类单元中。虽然总体微生物组成与 MUI 无关，但我们证实了某些细菌群落与 MUI 之间的关联。在评估多变量模型中细菌群落与 MUI 之间的关联时，纳入与尿微生物组相关的几个协变量可改善推论。
Zotero 插件
