Distinct microbiome composition of floor and air dust distort richness estimation in university dormitories

The number of culture-independent indoor microbiome study has increased remarkably in recent years, but microbial composition among different sampling strategies remains poorly characterized and their impact to downstream microbiome analysis is also not clear. In this study, we reported a case study of microbial composition of floor and air dust in 87 dormitory rooms of Shanxi University, China. Floor and air dust were collected by vacuum cleaner and petri-dish, respectively, and the bacterial composition was characterized by 16S rRNA sequencing. The composition of floor and air dust differed significantly (R2 = 0.65, p < 0.001, Adonis), and Pseudomonas dominated in floor dust (75.1%) but was less common in air dust (1.9%). The top genera in air dust, including Ralstonia (15.6%), Pelomonas (11.3%), Anoxybacillus (9.3%) and Ochrobactrum (6.2%), all accounted for < 1% abundance in floor dust. The dominant Pseudomonas in floor dust swamped low frequency organisms, leading to significant lower number of operational taxonomic units (OTUs) compared with air dust in the same sequencing depth. The different microbial composition of floor and air dust can lead to differences in downstream bioinformatics analyses. We searched the dormitory microbiome against ~200,000 samples deposited in Microbiome Search Engine (MSE), and found that the compositions of floor dust samples were similar to samples from building environment and human nasopharynx, whereas the compositions of the air dust samples were similar to mosquito tissues. ImportanceIncreasing number of indoor microbiome studies has been conducted in recent years, but the impact of sampling strategy is far from clear. In this study, we reported that the floor and air sampling can lead to drastic variation in microbial composition and downstream analyses in university dormitories, including microbial diversity estimation and compositional similarity search. Thus, the bioinformatics analysis results need to be interpreted with caution in microbiome studies, and it may be necessary to collected samples from different sampling strategies to comprehensively characterize the microbial composition and exposure to human occupants in an indoor environment.