Impact of industrial production system parameters on chicken microbiomes: mechanisms to improve performance and reduce Campylobacter

Abstract

Background: The factors affecting host-pathogen ecology in terms of the microbiome remain poorly studied. Chickens are a key source of protein with gut health heavily dependent on the complex microbiome which has key roles in nutrient assimilation and vitamin and amino acid biosynthesis. The chicken gut microbiome may be influenced by extrinsic production system parameters such as Placement Birds/m² (stocking density), feed type and additives. Such parameters, in addition to on-farm biosecurity may influence performance and also pathogenic bacterial numbers such as Campylobacter. In this study, three different production systems 'Normal' (N), 'Higher Welfare' (HW) and 'Omega-3 Higher Welfare' (O) were investigated in an industrial farm environment at day 7 and day 30 with a range of extrinsic parameters correlating performance with microbial dynamics and Campylobacter presence.

Results: Our data identified production system N as significantly dissimilar from production systems HW and O when comparing the prevalence of genera. An increase in Placement Birds/m² density led to a decrease in environmental pressure influencing the microbial community structure. Prevalence of genera, such as Eisenbergiella within HW and O, and likewise Alistipes within N were representative. These genera have roles directly relating to energy metabolism, amino acid, nucleotide and short chain fatty acid (SCFA) utilisation. Thus, an association exists between consistent and differentiating parameters of the production systems that affect feed utilisation, leading to competitive exclusion of genera based on competition for nutrients and other factors. Campylobacter was identified within specific production system and presence was linked with the increased diversity and increased environmental pressure on microbial community structure. Addition of Omega-3 though did alter prevalence of specific genera, in our analysis did not differentiate itself from HW production system. However, Omega-3 was linked with a positive impact on weight gain.

Conclusions: Overall, our results show that microbial communities in different industrial production systems are deterministic in elucidating the underlying biological confounders, and these recommendations are transferable to farm practices and diet manipulation leading to improved performance and better intervention strategies against Campylobacter within the food chain. Video Abstract.

Keywords: Campylobacter; Chicken; Competitive exclusion; Diversity; Environmental filtering; Microbiome; Phylogenetic signal; Production systems.

Fig. 1

Microbial diversity and community structure. a Alpha diversity (Richness and Shannon entropy) and environmental filtering (NRI/NTI) measures respectively. Lines (a) connect two categories where the differences were significant (ANOVA) with *P < 0.05, **P < 0.01, or ***p < 0.001. b Beta diversity using Bray-Curtis distance measure along with top-25 genera observed in all samples grouped by categories. The tables represent taxa that were found to be significant based on subset analysis (Supplementary S1), i.e. those genera selected in the subsets that explain roughly the same distance between samples as all the genera. Additionally, if the taxa were found to be differentially expressed based on other analyses, such as DESeq2 (Supplementary S11), MINT (Supplementary S2) and differential heat tree (Fig. 2), the categories they up- and downregulated are represented with corresponding up and down arrows. For example, in HW30 vs O30 comparison, ‘(S), O30 (D, H)’ for Phascolarctobacterium should be read as selected in subset analysis: (S) and upregulated in O30 according to both DESeq2 and Differential Tree: O30 (D, H)

Fig. 2

Taxa that persist and those that are differentially abundant. a Core microbiome analyses that persist in 85% of the samples for different production systems (day 7 and day 30). In the heat maps, the OTUs are sorted by their abundances with those on the top being low abundant, whereas those at the bottom are highly abundant. b Differential heat tree with taxonomy key given in the middle, and the branches where they are upregulated are coloured according to their respective categories shown on top of each subpanel

Fig. 3

Heatmap of key extrinsic parameters that influence different attributes of microbiome. The figure is based on subset regressions (Supplementary S4 to S10), where red and blue represent the significant positive and negative beta coefficients that were consistently selected in different regression models. The categorical variables are represented with a yellow highlight (coded as 1 (present) or 0 (absent)) and if selected is interpreted as the samples belonging to those categories having positive/negative influence on the respective microbiome metrics