Figure 3

Detection of single-cell expression of ruminant IL-17A by ELISpot. Plates and PBMC were prepared and cultured as described in “IL-17A ELISpot section”. ELISpot images shown are representative of PBMC from one of three cattle (A) and one of three sheep (B) activated with ConA and PMA/ionomycin. The average number of spot-forming units (SFU) with standard errors are shown for 10⁶ PBMC from all three cattle (grey bars) and sheep (black bars), stimulated under the different conditions (C). Data were modelled by fitting a Poisson generalised linear mixed model (GLMM) by maximum likelihood to the IL-17A SFU/10⁶ values, using logarithmic link function and Laplace approximations to calculate log-likelihoods. The model included treatment (medium control, ConA and PMA/ionomycin), species (bovine, ovine) and their interaction as fixed effects and animal identification as a random effect in order to account for both within- and between-animal variability. An observation-level random effect term was specified to account for data over-dispersion. The statistical significance of the fixed effect terms was assessed using p values derived from type II Wald Chi square tests. Linear hypothesis tests were defined from the GLMM in order to conduct pair-wise comparisons of means between treatments and species. The associated p values were adjusted for false discovery rate (FDR) following Benjamini–Hochberg’s procedure.