Table 3 Host covariates included in the variable selection process and retained for downstream CCA analysis.

From: Diet and gut microbiome enterotype are associated at the population level in African buffalo

Variable category

Included in envfit analyses

Included in CCA

Included in CCA with diet interaction term

Disease

Bovine tuberculosis (TB) status, Parainfluenza-3 incidence, Adenovirus-3 incidence, Mycoplasma bovis (MB) incidence, bovine respiratory syncytial virus (BRSV) incidence, bovine herpesvirus (BHV) status, bovine viral diarrhea virus (BVDV) status, strongyle burden (log transformed), coccidia burden (log transformed), trichuris burden (log transformed)

 

MB incidence, strongyle burden, BHV status, BVDV status

Nutrition

Body condition score (BCS), fecal chlorophyll, hematocrit (Hct), calcium, phosphorous, magnesium (MG), total protein

Hct

MG

  1. We performed envfit analysis on PCA objects generated from Bray–Curtis distance matrices within the restricted dietary regime, the hay/green vegetation regimes, and pooled samples. For each variable group, host covariates were considered “significant” within a given dataset if envfit analysis yielded p < 0.05. Permutations within the envfit algorithm were stratified by capture number to account for temporal effects. Chronic respiratory disease status (tuberculosis, bovine herpesvirus, bovine diarrhea virus) was coded as 0 if negative and 1 if positive. Acute respiratory infection incidence (parainfluenza-3, adenovirus-3, Mycoplasma bovis, bovine respiratory syncytial virus) were coded as 0 if they did not seroconvert and 1 if they did seroconvert between captures. Gastrointestinal parasite burdens were measured in eggs per gram of feces for each individual.
Back to article page