Generating super-shedders: co-infection increases bacterial load and egg production of a gastrointestinal helminth

Lass, Sandra; Hudson, Peter J.; Thakar, Juilee; Saric, Jasmina; Harvill, Eric; Albert, Réka; Perkins, Sarah E.

To determine whether the time course of bacterial load differed between single and co-infected treatment groups we used the model described by Fenton et al. [34] , where a generalized linear mixed model (GLMM) using ASReml v. 2.0 was used to determine differences in the number of faecal egg counts of multiple individuals over time. Mixed models allow the user to control for multiple variables at the same time, including both random and fixed effects. One advantage to ASReml is that spline terms are included in the random model where only 1 d.f. is used. Here, we fit this spline to the time course of infection in our data, allowing nonlinear relationships between variables to be modelled. We also included host identification as a random term to control for pseudo-replication (i.e. autocorrelation errors). For the fixed model, we fitted a spline to the bacterial load, over time, of the single and co-infected groups and used this as the response variable with treatment group (single or co-infected) as the explanatory variable. We first logtransformed the data and then carried out a GLMM analysis using ASReml in software R [35] . To assess whether co-infection altered host mortality, Cox proportional hazards were used to determine how survivorship differed between the treatment groups [35] .