Index

• aitchison_pca_based_analyses.Rmd: Aitchison PCA analysis
  • Figs. 2A and 3D; figs S3, S4, S10, and S12
• taxon_specific_linear_models_and_smooths.Rmd: Taxon-specific linear models and smooths
  • Fig. 2C; figs. S7, S8, and S9
• neutral_models.Rmd: Sloan's Neutral model
  • fig. S16
• autocorrelation_analysis.Rmd: Autocorrelation analysis
  • Figs. 3A, B, C; figs. S11
• taxon_log_ratio_analysis.Rmd: Host and social group log-ratios
  • Fig. 3E; figs. S14, S15, S20, and S21
• grooming_analysis.Rmd: Social connectedness - microbiome similarity analysis
  • figs. S23 and S24

All of the above analyses apart from grooming_analysis.Rmd: use the file ps.RDS in the OSF Data/ repository. grooming_analysis.Rmd: uses within_grp_groom.RDS in the OSF Data/ repository

• GAMs/
  • hierarchical/
    • gams_vi_pcs_and_alphadiv.R: Code for fitting model P, model P+G and model P+G+H to microbiome PC1-3 and 3 alpha diversity metrics
    • gams_vi_phylum.R: Code for fitting model P, model P+G and model P+G+H to the clr-transformed relative abundance of 12 phyla
    • gams_vi_family.R Code for fitting model P, model P+G and model P+G+H to the clr-transformed relative abundance of 34 families
  • variable_importance/
    • gams_all_models_PCs_div.R Code for fitting model P+G+H to microbiome PC1-3 and 3 alpha diversity metrics, successively removing one predictor variable at the time, keeping the model otherwise intact
    • gams_all_models_phylum.R Code for fitting model P+G+H to the clr-transformed relative abundance of 12 phyla, successively removing one predictor variable at the time, keeping the model otherwise intact
    • gams_all_models_family.R Code for fitting model P+G+H to the clr-transformed relative abundance of 34 families, successively removing one predictor variable at the time, keeping the model otherwise intact
  • santiy_checks/
    • randomize_host_and_grp_labels.Rmd: Code to randomize social group membership and host identity
    • gams_randm_grp_labels.R: Code to run model P+G+H with randomized social group labels
    • gams_randm_grp_labels.R: Code to run model P+G+H with randomized host labels
    • simulation.Rmd: Code to run a simulation investigating the effect of model complexity on deviance explained

All of the above GAMs use the file metadata_detrended.RDS in the OSF Data/ repository