Data and code from: Consistent, scale-dependent differences in the biogeography of host-associated and free-living microbiomes across systems
Data files
Jun 26, 2026 version files 12.10 MB
-
diff_abund_final.R
24.41 KB
-
dist_decay_final.R
32.51 KB
-
diversity_final.R
29.89 KB
-
metacomm_TSCC.csv
51.55 KB
-
physeq_filt_nonnormal.RDS
3.94 MB
-
physeq_normal.RDS
3.70 MB
-
physeq_srs_normalize.R
4.82 KB
-
PNM.R
52.34 KB
-
ps.tree.RDS
4.26 MB
-
README.md
3.40 KB
-
site_coords.csv
1.08 KB
-
site_data.csv
1.09 KB
Abstract
Microbial communities are critical to the functioning of ecosystems and shape the ecology and evolution of host organisms. Consequently, understanding their distribution, diversity, and composition across temporal and spatial scales is essential. However, we have a limited understanding of how host-associated and free-living microbes differ in their structure and biogeography. Here, we test whether host-associated (fish gut) and free-living (lake bacterioplankton) microbes exhibit different metacommunity structure, spatial turnover, and consistency with neutral expectations using two independent lake systems. We characterized microbial communities in lake water (Vancouver Island and Sierra Nevada) and guts in two fish species (stickleback and brook trout) using 16S amplicon sequencing. We compared alpha and beta-diversity within lakes, quantified spatial turnover (distance-decay), and tested for departure from neutral abundance-occurrence expectations between bacterioplankton and fish gut microbiomes. Fish microbiomes had lower alpha-diversity compared to bacterioplankton, but higher beta-diversity within lakes. Bacterioplankton were more similar across lakes, yet showed stronger patterns of spatial turnover with distance than fish gut microbiomes. A neutral model explained a substantial proportion of abundance-occurrence relationships in bacterioplankton communities but performed poorly for fish-associated microbes. Our study indicates that host-associated and free-living microbes have disparate patterns of metacommunity structure and spatial turnover consistent with differences in the strength of neutral ecological processes. Fish microbiomes were less diverse at the local scale but more variable across space and time than bacterioplankton communities, suggestive of potentially strong local selection and/or reduced microbial exchange among hosts compared to environmental communities. Importantly, we observed highly consistent patterns across both lake systems despite differences in host species, sampling design, and region, demonstrating that differences in the distribution of host and environmental microbes are potentially widespread. Our results demonstrate how host association fundamentally alters the diversity and spatial distribution of microbes, emphasizing the need to incorporate hosts into broader frameworks of microbial biogeography.
We have a limited understanding of how host-associated and free-living microbes vary in their distribution, diversity, and composition. Here, we test whether host-associated (fish gut) and free-living (lake water) bacteria and archaea have different metacommunity structure and biogeography across two distinct lake ecosystems.
File Directory
The file directory contains folders housing data (csvs and R objects) and scripts (R) for analysis.
Scripts
File: diff_abund_final.R
Description:
script for performing differential abundance analysis and taxa barplots for comparisons of host-associated and free-living microbiomes. File: dist_decay_final.R Description: script for performing distance decay (spatial autocorrelation) analysis.
File: diversity_final.R
Description:
script for calculating and comparing alpha and beta-diversity metrics between host-associated and free-living microbiomes.
File: ordinations_final.R
Description:
script for calculating ordinations (Principle Coordinates Analysis) and PERMANOVA analysis to examine factors explaining variation in community structure.
File: physeq_srs_normalize.R
Description:
script for preprocessing of phyloseq object output downstream of DADA2 pipeline. Includes Scaling with Ranked Subsampling (SRS) to normalize read counts to account for differences in read depth among samples.
File: PNM.R
Description:
script for fitting the Prokaryote Neutral Model to sample categories.
Data
File: metacomm_TSCC.csv
Description:
csv file containing all samples corresponding to fastq sequence files used in analysis.
Variables:
- sample_id: file and sample names
- sample_type: represents whether samples are stickleback, Vancouver Island water, Sierra Nevada water, or brook trout
- habitat_type: indicates whether samples are stickleback, water, or brook trout
- site: represents lakes where samples were collected
- watershed: indicates basins where samples were collected
- sex: represents the sex of fish where applicable
- year: represents the year where samples were collected
- run_no: represents the MISEQ run where samples were sequenced (this applies only to the Sierra Nevada)
File: ps.tree.RDS
Description:
RDS file of phyloseq object before filtering out unwanted samples and normalization (SRS).
File: physeq_normal.RDS
Description:
RDS file of phyloseq object after filtering and normalization.
File: site_coords.csv
Description:
csv file of coordinates of lakes samples.
Variables:
- site: the lake where samples were collected
- longitude: longitude of sites in degrees
- latitude: latitude of sites in degrees
File: site_data.csv
Description:
csv file of coordinates and elevation data of lakes sampled.
Variables:
- site: lakes in the Sierra Nevada
- longitude: longitude of sites in degrees
- latitude: latitude of sites in degrees
- elevation: elevation in meters where lakes are located
- X: coordinates of sites for mapping on the x plane
- Y: coordinates of sites for mapping on the y plane
File: physeq_filt_nonnormal.RDS
Description:
RDS file of phyloseq object after filtering out unwanted/unanalyzed samples but before SRS normalization.
