Scholz C, Jarquín-Díaz VH, Planillo A, Radchuk V, Scherer C, Schulze C, Ortmann S, Kramer-Schadt S, Heitlinger E (2024): Host weight, seasonality and anthropogenic factors contribute to parasite community differences between urban and rural foxes Host weight, seasonality and anthropogenic factors contribute to parasite community differences between urban and rural foxes. STOTEN, 936. doi:10.1016/j.scitotenv.2024.173355
Pathogens often occur at different prevalence along environmental gradients. This is of particular importance for gradients of anthropogenic impact such as rural-urban transitions presenting a changing interface between humans and wildlife. The assembly of parasite communities is affected by both the external environmental conditions and individual host characteristics. Hosts with low body weight (smaller individuals or animals with poor body condition) might be more susceptible to infection. Furthermore, parasites' mode of transmission might affect their occurrence: rural environments with better availability of intermediate hosts might favour trophic transmission, while urban environments, typically with dense definitive host populations, might favour direct transmission. We here study helminth communities (141 intestinal samples) within the red fox (Vulpes vulpes), a synanthropic host, using DNA metabarcoding of multiple marker genes. We analysed the effect of urbanisation, seasonality and host-intrinsic (weight, sex) variables on helminth communities. Helminth species richness increased in foxes with lower body weight and in winter and spring. Season and urbanisation, however, had strong effects on the community composition, i.e., on the identity of the detected species. Surprisingly, transmission in two-host life cycles (trophic transmission) was more pronounced in urban Berlin than in rural Brandenburg. This disagrees with the prevailing hypothesis that trophically transmitted helminths are less prevalent in urban areas than in rural areas. Generally, co-infestations with multiple helminths and high infection intensity are associated with lighter (younger, smaller or low body condition) animals. Both host-intrinsic traits and environmental drivers together shape parasite community composition and turnover along urban-rural gradients.
This readme describes the analysis pipeline in terms of a data-processing pipeline: where raw and intermediate data is stored and processed. For more insight into the methods used consult the manuscript text and the R code.
For a fully reproducible analysis starting from raw sequencing files R script have to be executed in numerical order (starting with a "\d_"; e.g. 0_, 1_, ..).
We follow one simple convention for executing the scripts: Your R session should run in the repository folder directly (in the folder that you cloned, not in 'input_data/', '/intermediate_data' or 'R/').
At the beginning of each script we test for the availability of the required data (in a potentially interactive R session) and either recompute it executing ('sourcing') the previous scripts (recompute = TRUE) or read it from 'intermediate_data'.
All neccesary intermediate data to run the script are contained within this repository. You can run each script seperately to scrutinze the methods and code an to reproduce the results.
| File reference | Created in | File in ropository |
|---|---|---|
| Figure 1 | R/1_Fox_general_MA.R | figures/map_study_overview_multi.png |
| suppl. table 1 | outside, manually | input_data/primer_file_foxes.csv |
| Table 1 | outside, manually | text only |
| suppl. figure 1 | R/2_iNEXT_fox.R | figures/suppl/CorrelatPedictors.png |
| suppl. table 2 | outside, manually | input_data/helminth_traits.csv |
| Tabe2 | R/2_iNEXT_fox.R | tables/prevalences.html |
| Figure 3 | R/2_iNEXT_fox.R | figures/Div_Model.png |
| Suppl. figure 2 | R/2_iNEXT_fox.R | figures/suppl/NumberSeqVar.png |
| Suppl. figure 3 | R/2_iNEXT_fox.R | figures/suppl/conditionVar.png |
| Figure 4 | R/3_compositionHelm.R | figures/CompositionEnvHelm.png |
| Table 3 | R/3_compositionHelm.R | tables/Permanova.csv |
| suppl. table 4 | R/3_compositionHelm.R | tables/PermanovaConti.csv |
| Figure 5 | R/4b_JSDM_helmAnalysis.R | figures/PAModel_area_varpart.png |
| Figure 6 | R/4b_JSDM_helmAnalysis.R | figures/PAModel_area_BetaCoefs.png |
| Figure 7 | R/4b_JSDM_helmAnalysis.R | figures/PAModel_area_GammaCoefs_traits.png |
| suppl figure 4 | R/4b_JSDM_helmAnalysis.R | figures/suppl/VarPart_PAModel_grad.png |
| Suppl.figure 5 | R/4b_JSDM_helmAnalysis.R | figures/suppl/PAModel_grad_BetaCoefs.png |
| Suppl.figure 6 | R/4b_JSDM_helmAnalysis.R | figures/suppl/PAModel_grad_GammaCoefs_traits.png |
-> R/0_Extract_Einvir_Covariates.R (by Cedric Scherer and Aimara Planillo)
The input raw (layer) files this is based on are in: input_data/tifs/*.tif
As this contains only impervious surface, tree cover and human footprint index, unlike data for previous (pre-mid-2021) versions of the manuscript, now allows us analysis of landscape variables for both Berlin and Brandenburg.
The script reads data on the sampled foxes from "input_data/Fox_data.csv', appends environmental variables for each fox and writes them (together with the 'basic data') to 'intermediate_data/Fox_data_envir.RDS'.
-> R/1_Fox_general_MA.R (by Victor Victor Jarquin-Diaz and Emanuel Heitlinger)
The script reads the raw sequencing data (and stors and reads intermediate files) from the compute server of the Heitlinger group ([email protected]) and is in the present form not executable on other systems. In oder to execute it you'll need to download the data from NCBI-SRA BioProject PRJNA386767, as sequecing data is to large for storage on github. Then adapt the path for reading the data in the script accoringly.
We process the raw sequencing data based on matching of the primer sequences in 'input_data/primer_file_foxes.csv'. We are using the MultiAmplicon wrapper of the dada2 package to produce amplified sequence variant (ASV) abundances for each fox.
The script also adds the environmenal covariates (from intermediate_data/Fox_data_envir.RDS produced in 0_Extract_Einvir_Covariates.R) to the central "phyloseq" object of the pipeline. The environmental covariates are stored as "sample_data" in this object.
Similarly, the helminth trait data (found in input_data/helminth_traits.csv; compiled by Carolin Scholz and Emanuel Heitlinger) is storted in the "phloseq" object's "taxon_table".
We store output as a phyloseq object in 'intermediate_data/PhyloSeqCombi.Rds'
-> R/2_iNEXT_fox.R (by Carolin Scholz, Aimara Planillo, Cedric Scherer and Emanuel Heitlinger)
We look at three differen masures for alpha diverstiy: species richness [q=0], Shannon diversity [q=1] and Simpson diversity [q=2]), based on the iNext package. We compare beta diversity (jaccard centroid distances) using the package vegan and gamma diversity (rarefied diversity by fox individual) again using iNext.
Diversity data for each fox are appended to the "sample_data" in the phyloseq object.
-> 3_CommunityComposition.R (by Emanuel Heitlinger)
Performs community analyses using the vegan package
-> 4a_JSDM_helminths.R (by Aimara Planillo) Runs the JSDM models
-> 4b_JSDM_helmAnalysis.R Evaluates the models and extract posterior effects and plot in