Fig. 5

Chicken intestinal microbiota analyses. A Heatmap of the 50 most abundant phylotypes of 156 samples collected. Five types of intestinal microbiota were revealed based on clustering patterns. Ward linkage clustering is used to cluster samples based on their Jensen-Shannon distance calculated in vegan package in R [116]. Identified microbiota types are labeled as I–V. B Distribution of different community type by cluster as shown in A in diet treatment groups of Control, colostrum and NFDM. Type I is significantly lowered in control than in other two groups while type IV is significantly higher in BC and NFDM. Type II is borderline significantly higher in BC and NFDM than control, based on statistical modeling using Bayesian Poisson model. C Phylotype biomarkers for digesta and scrapings generated using program LEfSe [105]. Vertical bars represent the relative abundance of Bacteriodes in each sample. Dotted line represents mean, solid line represents median relative abundance. The alpha value for the non-parametric factorial Kruskal–Wallis (KW) sum-rank test was 0.05 and the threshold for the logarithmic LDA model (43) score for discriminative features was set at 2.0