Co-occurrence patterns of bacteria within microbiome of Moscow subway

Abstract

Microbial ecosystems of the built environments have become key mediators of health as people worldwide tend to spend large amount of time indoors. Underexposure to microbes at an early age is linked to increased risks of allergic and autoimmune diseases. Transportation systems are of particular interest, as they are globally the largest space for interactions between city-dwellers. Here we performed the first pilot study of the Moscow subway microbiome by analyzing swabs collected from 5 types of surfaces at 4 stations using high-throughput 16S rRNA gene sequencing. The study was conducted as a part of The Metagenomics and Metadesign of the Subways and Urban Biomes (MetaSUB) project. The most abundant microbial taxa comprising the subway microbiome originated from soil and human skin. Microbiome diversity was positively correlated with passenger traffic. No substantial evidence of major human pathogens presence was found. Co-occurrence analysis revealed clusters of microbial genera including combinations of microbes likely originating from different niches. The clusters as well as the most abundant microbes were similar to ones obtained for the published data on New-York City subway microbiome. Our results suggest that people are the main source and driving force of diversity in subway-associated microbiome. The data form a basis for a wider survey of Moscow subway microbiome to explore its longitudinal dynamics by analyzing an extended set of sample types and stations. Complementation of methods with viability testing, "shotgun" metagenomics, sequencing of bacterial isolates and culturomics will provide insights for public health, biosafety, microbial ecology and urban design.

Keywords: 16S rRNA; Biosurveillance; Built environments; Co-occurrence patterns; Subway; Urban microbiome.

Fig. 1

Top ten most prevalent microbial genera for different subway stations and sample types. Barplot height corresponds to the average genus abundance for all samples from a specific group. In the left and right columns of barplots, colors represent the types of surfaces and stations, respectively (see legend on the right). The “NNN_u” notation refers to the “unclassified genus(-era) from family NNN”.

Fig. 2

Variation of alpha-diversity (Chao1 diversity index) across stations and surface types.

Fig. 3

Association between daily passenger traffic and alpha-diversity. The four stacks of samples correspond to the stations in the following order: Vystavochnaya, Sretenskiy boulevard, Dostoyevskaya and Rimskaya.

Fig. 4

Distribution of microbial community structures across Moscow subway stations and sample types using two metrics (principal coordinates analysis, PCoA): Bray-Curtis metric (A) and Jaccard metric (B). Arrows in the bi-plots show major microbial drivers of the observed variance.

Fig. 5

UpSet plots illustrating quantitative intersection of the sets of microbial genera across the stations (A) and surfaces (B). The numbers above the bars show the number of common genera between the groups of samples marked below the bars.

Fig. 6

Microbial clusters (co-occurrence groups of species) calculated using SPIEC-EASI. Vertex diameter is proportional to the abundance of the taxon. The taxa likely originating from human microbiome are shown in pink. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)