Skip to main content
Dryad

Community metabarcoding reveals the relative role of environmental filtering and spatial processes in metacommunity dynamics of soil microarthropods across a mosaic of montane forests

Cite this dataset

Noguerales, Víctor et al. (2021). Community metabarcoding reveals the relative role of environmental filtering and spatial processes in metacommunity dynamics of soil microarthropods across a mosaic of montane forests [Dataset]. Dryad. https://doi.org/10.5061/dryad.s4mw6m97c

Abstract

Disentangling the relative role of environmental filtering and spatial processes in driving metacommunity structure across mountainous regions remains challenging, as the way we quantify spatial connectivity in topographically and environmentally heterogeneous landscapes can influence our perception of which process predominates. More empirical datasets are required to account for taxon- and context-dependency but relevant research in understudied areas is often compromised by the taxonomic impedimentWe here employed haplotype-level community DNA metabarcoding, enabled by stringent filtering of Amplicon Sequence Variants (ASVs), to characterize metacommunity structure of soil microarthropod assemblages across a mosaic of five forest habitats on the Troodos mountain range in Cyprus. We found similar β diversity patterns at ASV and species (OTU, Operational Taxonomic Unit) levels, which pointed to a primary role of habitat filtering resulting in the existence of largely distinct metacommunities linked to different forest types. Within-habitat turnover was correlated to topoclimatic heterogeneity, again emphasizing the role of environmental filtering. However, when integrating landscape matrix information for the highly fragmented Quercus alnifolia habitat, we also detected a major role of spatial isolation determined by patch connectivity, indicating that stochastic and niche-based processes synergistically govern community assembly. Alpha diversity patterns varied between ASV and OTU levels, with OTU richness decreasing with elevation and ASV richness following a longitudinal gradient, potentially reflecting a decline of genetic diversity eastwards due to historical pressures. Our study demonstrates the utility of haplotype-level community metabarcoding for characterising metacommunity structure of complex assemblages and improving our understanding of biodiversity dynamics across mountainous landscapes worldwide.

Usage notes

Raster layers
 
This ZIP folder contains the TIF files corresponding to the digital elevation model (“01_DEM” subfolder), six WorldClim layers (“02_WorldClim” subfolder), four ENVIREM layers (“03_ENVIREM” subfolder) and the spatial distribution of Quercus alnifolia patches (“04_Quercus_alnifolia” subfolder). This latter subfolder also contains input files (ASCII format) for CIRCUITSCAPE corresponding to the isolation-by-resistance (IBR) scenarios defined by (i) the distribution of Quercus alnifolia forest patches assuming increasing resistance values (from 5 to 1,000,000) for non-Quercus cells, (ii) the topographic complexity as estimated by the terrain roughness index (TRI) and (iii) a “flat” scenario (NULL) whose cells have a fixed resistance (=1) value. Geographic coordinates of Quercus alnifolia sampling sites are also provided in the format required by CIRCUITSCAPE. The WorldClim and ENVIREM layers were processed and interpolated at 90 meters resolution as detailed in Supplemental Information.

File: 0A_Raster_layers.zip

Community tables
 
Filtered community tables (taxa in rows, sites in columns) used for downstream analyses, at both ASV and OTU levels. Sampling site codes as in Table S1 in Supplemental Information.

File: 0B_Community_tables.zip

 
DNA sequences
 
Files in FASTA format containing DNA sequences of (a) fully-filtered ASVs derived from metabarcoding data and (b) Sanger sequences of ‘voucher’ specimens.

File: 0C_DNAsequences.zip

 
GLMM input file
 
Input file for Generalized Linear Mixed Models (GLMM), which were used to analyse the relationship between ASV- and OTU-based richness (RICH) or local contribution to beta diversity (LCBD) per site and the topoclimatic variables (ENVPC1 and ENVPC2) as predictors, with latitude and longitude as covariates and forest habitat type as a random effect. Sampling site codes as in Table S1 in Supplemental Information.

File: 01_GLMM.zip

 
PCNM input files
 
This ZIP folder contains the topographic weighted distance matrices (SPATWD) used as input files in Principal Coordinates of Neighbour Matrices (PCNM) analyses to generate spatial predictors, at across-habitats and within-habitat scales. Sampling site codes as in Table S1 in Supplemental Information.

File: 02_PCNM.zip


dbRDA input files
 
This ZIP folder contains the input files used to perform Distance-Based Redundancy Analysis (dbRDA), at both across-habitats and within-habitat scales, including (a) the community dissimilarity matrices based on the Simpson dissimilarity index calculated at both ASV and OTU level, and (b) tables of spatial (SPAPCNMi) and environmental predictors (habitat type, ENVPC1 and ENVPC2). Sampling site codes as in Table S1 in Supplemental Information.
 
File: 03_dbRDA.zip

 
mvGLM input files

This ZIP folder contains the input files used to perform Multivariate Generalized Linear Models (mvGLMs), at both across-habitats and within-habitat scales, including (a) community tables (presence/absence) for both ASV and OTU level, and (b) spatial (SPAPCNMi) and environmental predictors (habitat type, ENVPC1 and ENVPC2). Sampling site codes as in Table S1 in Supplemental Information.
 
File: 04_mvGLM.zip

 
MMRR and distance decay input files
 
This ZIP folder contains the input files used to perform Multivariate Matrix Regression with Randomization (MMRR), including (a) community dissimilarity matrices based on the Simpson dissimilarity index calculated at both ASV and OTU level for Quercus alnifolia sampling sites, and (b) distance matrices based on (i) resistance due to habitat fragmentation (FRAIBR) assuming increasing resistance values (from 5 to 1,000,000) for non-Quercus cells, (ii) resistance due to topographic complexity (TRIIBR), (iii) resistance due to a “flat landscape” (NULLIBR), (iv) weighted topographic (SPATWD) distances and (v) topoclimatic (ENVPC1-2) distances. Sampling site codes as in Table S1 in Supplemental Information.

File: 05_MMRR-DistanceDecay.zip