Additional file 2 of Bacterial colonization dynamics and antibiotic resistance gene dissemination in the hospital environment after first patient occupancy: a longitudinal metagenetic study

Additional file 1: Suppl. Figure S1. Workflow schematic of the longitudinal study. Shown is the map depicting the 9 rooms of the neurological ward in which 1547 samples were collected including 3 environmental sites and 4 patient sites over a time course of 31 weeks (Pre-opening week + 30 weeks after initial patient occupancy). The metagenetic pipelines included 16S rRNA quantification and sequencing steps, as well as Taq-Man assays for ARG-detection. Suppl. Figure S2. Correlation between physical parameters and bacterial load. Shown are different multiple regression models which beside location include either the temperature, the humidity or both factors as potential contributors to the bacterial load as measured by 16S rRNA copies (qPCR). The highest r2 value and the best model fit was achieved when combining both temperature and humidity in the regression model. (C/E: copies per extraction). Suppl. Figure S3. Alpha- and beta-diversity metrics used for systematic subsampling in week-blocks. A) Shown are the overall alpha-diversity values of the collapsed environmental microbiome of each of the 31 weeks. Significant increase was observed after the first initial 4-week block. B) PCA of the collapsed environmental microbiome data for all 31 week (each week represented by a single dot). The coloring scheme is based on the distance pattern between consecutive weeks. The second block (8 weeks, green) was further divided in two halves to allow a more detailed analysis of the critical period of community stabilization, which was shown to occur between weeks 5 and 7 after patient occupancy. Suppl. Figure S4. Alpha-diversity metrics of the different room sites over time. Shown are the Shannon indices (Median and IQR 25-75%) of floor, doorhandle and sink samples across different week blocks. Suppl. Figure S5. Distance comparison between sites in the pre-opening and last sampling weeks. Shown are the weighted UniFrac distances between site pairs as measured in week 0 and in week 30 (Mean±SEM; *p<0.05; **p<0.01). Suppl. Figure S6. No impact of spatial room distribution on microbiome patterns. A) Principal Coordinate plots show the distribution of the microbiome data based on the distribution of the rooms across the ward. Shown are the p-values obtained by PERMANOVA test. B) The floor plan shows two clusters of contiguous rooms among the sampled sites, which are located on opposing sides of the neurology ward. Suppl. Figure S7. Patient microbiome patterns in this study. A) Collapsed taxonomic summary of the nasal and rectal swabs collected from the patients during the first 30 weeks of occupancy. B) Heatmaps showing the amount of shared taxa (at family level) between the core microbiomes of the environmental- and the patient-samples for different week-blocks. C) Taxonomic summary of the hand and elbow samples collected during the final 6 weeks of the time series. Suppl. Figure S8. Correlation plots between alpha-diversity metrics and external factors (temperature and humidity). Shown are the correlations between the temperature (°C) or the humidity (%) and the alpha-diversity of the three environmental sites. Suppl. Figure S9. PCoA of the beta-diversity depicting the distances between environmental and patient samples. Shown are the weighted UniFrac distances between the environmental cluster (grey) to each of the patient sites individually (colored). Suppl. Figure S10. Dynamics of the bacterial community structure over time. A) Relative abundance changes over time of the most abundant families. Bubble size represents the proportional abundance across different week blocks. B) Relative abundance dynamics of the top significantly changed taxa (family level) across the different week blocks. Suppl. Figure S11. Pathogen colonization of the hospital environment. A) Bar-chart depicting the incidence of pathogenic bacteria in the analyzed environmental samples. Shown are the number of samples positive for C. difficile or any of the 10 most frequent pathogens isolated from Charité-patients during the sampling period. B) Relative abundance of the detected pathogen sequences over time. Shown are the patterns across different week blocks. Suppl. Figure S12. ARG detection in the patient samples (rectum and nare) over selected weeks. Chart depicts the percentage of positive samples in each week for each of the ARGs. Suppl. Table S1. Basic epidemiological parameters of the patients sampled in this study. Suppl. Table S2. Library PCR-constructs. Shown are the sequences of the Fw- and Rv-primers used for library construction. Suppl. Table S3. List of nodes from the network analysis, indicating their pathogenicity status and the supporting reference source. When the taxa was not listed in any of the pathogen databases (KEGG Pathogens database ( https://www.kegg.jp/kegg/genome/pathogen.html ); ISID: Database of The International Society for Infectious Diseases (ISID) ( https://isid.org/ )), literature research was performed with the species name. Reports on opportunistic pathogenicity were referenced where possible. In all other cases, the node was defined as non-pathogen/unclassified. Suppl. Table S4. List of all ARGs addressed in this study, indicating the sequences of the designed primers and probes used for their detection via Taq-Man assay. In all cases, primers were designed to cover maximum number of known gene variants. Suppl. Table S5. Correlation tests between the occurrence of specific ARGs in Patient AND Environment Sites in a particular week. Shown are the p-values obtained for each correlation (Chi-Square test).