Average (mean) daily weight gain (ADG) and feed conversions (F:G; ratio of feed weight to gained weight of cattle) were calculated as: ADG=Total weight gain of cattle (as defined below)Total cattle days F:G=Total dry matter weight of feedTotal weight gain of cattle (as defined below)where total weight gain of cattle equals out-weight of cattle finishing the trial plus out-weight of cattle culled plus out-weight of dead cattle minus total enrollment weight

of cattle. Feedlot personnel performed daily health monitoring following standardized procedures. Animals were weighed individually at the beginning and end of the study. Fresh fecal samples (30/pen) from animals observed defecating were collected from separate pats in multiple areas throughout the pen. Care was taken to avoid ground contamination. Pens were Everolimus sampled weekly for four consecutive weeks prior to study end-dates for each block. Samples (approximately 30 g) were placed in sterile bags, stored in coolers, and transported to KSU for refrigeration (4 °C) until the following morning. Samples were cultured for E. coli O157:H7 using IMS and direct plating methods previously described [7] and [8]. Confirmation included a multiplex PCR for identifying the rfbE (O157), eae (intimin), stx1 (Shiga toxin 1), stx2 (Shiga toxin 2), hlyA (hemolysin),

and fliC (H7) genes [17]. Pen-level general and generalized linear mixed models (LMM and GLMM, respectively) NVP-BKM120 datasheet were used to assess potential treatment effects. For response variables recorded as pen-level proportions, data were fit using a GLMM with a binomial distribution and a logit link. Prevalence outcomes were the proportion of

samples positive of the total samples collected within the pen at each sampling. Mortality and culling risks were proportions based on the number of animals that died or were culled, respectively, during the study period out of the total number of animals enrolled within the pen. Data on ADG and F:G were modeled using LMM that assume a Gaussian distribution. For all models, random effects were fitted to recognize block as the clustering factor and pen as the experimental unit for treatment. For E. coli data, additional random effects were used to account for pen-specific repeated Farnesyltransferase measures over time. Independent variables included treatments (VAC, DFM, VAC x DFM interaction), and for E. coli data, effects of time and time-by-treatment interaction. Model diagnostics were based on studentized residuals (LMM) and functions of the Pearson χ2 statistic (GLMM). P values <0.05 were considered significant. Model-adjusted means (lsmeans back transformed to original scale) and SE were reported, and used to estimate vaccine efficacy using standard formula [18]. Study pens were filled with 17,148 steers. Pen sizes ranged between 398 and 464 steers (mean = 430.0). Mean weight at enrollment was 378.