Pascal Andreu Victória, Augustijn Hannah E, van den Berg Koen, van der Hooft Justin J J, Fischbach Michael A, Medema Marnix H
Bioinformatics Group, Wageningen University, Wageningen, the Netherlands.
Department of Bioengineering, Stanford University, Stanford, California, USA.
mSystems. 2021 Oct 26;6(5):e0093721. doi: 10.1128/mSystems.00937-21. Epub 2021 Sep 28.
Microbial gene clusters encoding the biosynthesis of primary and secondary metabolites play key roles in shaping microbial ecosystems and driving microbiome-associated phenotypes. Although effective approaches exist to evaluate the metabolic potential of such bacteria through identification of these metabolic gene clusters in their genomes, no automated pipelines exist to profile the abundance and expression levels of such gene clusters in microbiome samples to generate hypotheses about their functional roles, and to find associations with phenotypes of interest. Here, we describe BiG-MAP, a bioinformatic tool to profile abundance and expression levels of gene clusters across metagenomic and metatranscriptomic data and evaluate their differential abundance and expression under different conditions. To illustrate its usefulness, we analyzed 96 metagenomic samples from healthy and caries-associated human oral microbiome samples and identified 252 gene clusters, including unreported ones, that were significantly more abundant in either phenotype. Among them, we found the operon, a gene cluster known to be associated with tooth decay. Additionally, we found a putative reuterin biosynthetic gene cluster from a Streptococcus strain to be enriched but not exclusively found in healthy samples; metabolomic data from the same samples showed masses with fragmentation patterns consistent with (poly)acrolein, which is known to spontaneously form from the products of the reuterin pathway and has been previously shown to inhibit pathogenic Streptococcus mutans strains. Thus, we show how BiG-MAP can be used to generate new hypotheses on potential drivers of microbiome-associated phenotypes and prioritize the experimental characterization of relevant gene clusters that may mediate them. Microbes play an increasingly recognized role in determining host-associated phenotypes by producing small molecules that interact with other microorganisms or host cells. The production of these molecules is often encoded in syntenic genomic regions, also known as gene clusters. With the increasing numbers of (multi)omics data sets that can help in understanding complex ecosystems at a much deeper level, there is a need to create tools that can automate the process of analyzing these gene clusters across omics data sets. This report presents a new software tool called BiG-MAP, which allows assessing gene cluster abundance and expression in microbiome samples using metagenomic and metatranscriptomic data. Here, we describe the tool and its functionalities, as well as its validation using a mock community. Finally, using an oral microbiome data set, we show how it can be used to generate hypotheses regarding the functional roles of gene clusters in mediating host phenotypes.
编码初级和次级代谢产物生物合成的微生物基因簇在塑造微生物生态系统和驱动微生物组相关表型方面发挥着关键作用。尽管存在通过鉴定细菌基因组中的这些代谢基因簇来评估其代谢潜力的有效方法，但目前还没有自动化流程来分析微生物组样本中此类基因簇的丰度和表达水平，以生成关于其功能作用的假设，并找到与感兴趣表型的关联。在这里，我们描述了BiG-MAP，这是一种生物信息学工具，用于分析宏基因组和宏转录组数据中基因簇的丰度和表达水平，并评估它们在不同条件下的差异丰度和表达。为了说明其有用性，我们分析了来自健康和龋齿相关人类口腔微生物组样本的96个宏基因组样本，鉴定出252个基因簇，包括未报道的基因簇，它们在任一表型中都显著更丰富。其中，我们发现了与龋齿相关的操纵子基因簇。此外，我们发现来自一株链球菌的一个假定的罗伊氏菌素生物合成基因簇在健康样本中富集但并非仅在健康样本中存在；来自相同样本的代谢组学数据显示，具有与（聚）丙烯醛一致的碎片化模式的质量峰，已知（聚）丙烯醛可由罗伊氏菌素途径的产物自发形成，并且先前已证明其可抑制致病性变形链球菌菌株。因此，我们展示了BiG-MAP如何用于生成关于微生物组相关表型潜在驱动因素的新假设，并优先对可能介导这些表型的相关基因簇进行实验表征。微生物通过产生与其他微生物或宿主细胞相互作用的小分子，在决定宿主相关表型方面发挥着越来越被认可的作用。这些分子的产生通常由共线性基因组区域编码，也称为基因簇。随着越来越多的（多）组学数据集有助于在更深层次上理解复杂生态系统，需要创建能够自动分析跨组学数据集的这些基因簇的工具。本报告介绍了一种名为BiG-MAP的新软件工具，它允许使用宏基因组和宏转录组数据评估微生物组样本中基因簇的丰度和表达。在这里，我们描述了该工具及其功能，以及使用模拟群落对其进行的验证。最后，使用口腔微生物组数据集，我们展示了它如何用于生成关于基因簇在介导宿主表型中的功能作用的假设。
