Large-scale analysis of sheep rumen metagenome profiles captured by reduced representation sequencing reveals individual profiles are influenced by the environment and genetics of the host

Abstract

Background Producing animal protein while reducing the animal’s impact on the environment, e.g., through improved feed efficiency and lowered methane emissions, has gained interest in recent years. Genetic selection is one possible path to reduce the environmental impact of livestock production, but these traits are difficult and expensive to measure on many animals. The rumen microbiome may serve as a proxy for these traits due to its role in feed digestion. Restriction enzyme-reduced representation sequencing (RE-RRS) is a high-throughput and cost-effective approach to rumen metagenome profiling, but the systematic (e.g., sequencing) and biological factors influencing the resulting reference based (RB) and reference free (RF) profiles need to be explored before widespread industry adoption is possible. Results Metagenome profiles were generated by RE-RRS of 4,479 rumen samples collected from 1,708 sheep, and assigned to eight groups based on diet, age, time off feed, and country (New Zealand or Australia) at the time of sample collection. Systematic effects were found to have minimal influence on metagenome profiles. Diet was a major driver of differences between samples, followed by time off feed, then age of the sheep. The RF approach resulted in more reads being assigned per sample and afforded greater resolution when distinguishing between groups than the RB approach. Normalizing relative abundances within the sampling Cohort abolished structures related to age, diet, and time off feed, allowing a clear signal based on methane emissions to be elucidated. Genus-level abundances of rumen microbes showed low-to-moderate heritability and repeatability and were consistent between diets. Conclusions Variation in rumen metagenomic profiles was influenced by diet, age, time off feed and genetics. Not accounting for environmental factors may limit the ability to associate the profile with traits of interest. However, these differences can be accounted for by adjusting for Cohort effects, revealing robust biological signals. The abundances of some genera were consistently heritable and repeatable across different environments, suggesting that metagenomic profiles could be used to predict an individual’s future performance, or performance of its offspring, in a range of environments. These results highlight the potential of using rumen metagenomic profiles for selection purposes in a practical, agricultural setting.

Combining host and rumen metagenome profiling for selection in sheep: prediction of methane, feed efficiency, production, and health traits

“…Animals were born between 2014 and 2016 from three New Zealand ocks: one AgResearch ock (Flock 2638), one selection line ock containing sheep selected for high or low methane yield [18] (Flock 3633), and one Central Progeny Test ock (Flock 4640). Metagenome pro les were generated from rumen samples that were collected from the sheep at various time points throughout their life [14]. The metagenome pro les included in this study (Table 1) included samples taken when the sheep were fed two different diets: lucerne pellets (Lucerne) and New Zealand ryegrass-based pasture (Grass); and were classi ed as two age groups: Lambs (less than 15 months old) and Adults (greater than 15 months old).…”

“…The metagenome pro les included in this study (Table 1) included samples taken when the sheep were fed two different diets: lucerne pellets (Lucerne) and New Zealand ryegrass-based pasture (Grass); and were classi ed as two age groups: Lambs (less than 15 months old) and Adults (greater than 15 months old). The combination of feed and age resulted in three groups that were evaluated separately in this study: Grass Lamb, Grass Adult and Lucerne Lamb, corresponding to the GLS, GAS and LLS Groups in Hess et al [14]. All rumen samples were taken from animals when they were fed ad libitum and had been off feed for 2-4 hours (i.e.…”

“…All rumen samples were taken from animals when they were fed ad libitum and had been off feed for 2-4 hours (i.e. "Short" time off feed in [14]). Across the 1,200 sheep, 118 sires were represented, with each sire having 10 ± 5 offspring.…”

“…It has been shown that speci c microbes, as well as the entire microbial pro le, can be associated with both greenhouse gas emissions and feed e ciency [6-10], with a signi cant proportion of phenotypic variation explained by relationships between individual rumen microbiomes, often referred to as microbiability [11]. Furthermore, these microbes and pro les have been shown to be heritable [12][13][14], suggesting sustained progress can be achieved through selection practices. However, previous studies have typically been on a small scale and used technologies that either do not scale well when considering implementation in industry (e.g., whole genome sequencing) or do not capture the breadth of diversity in the rumen (i.e., prokaryotic 16S rRNA gene sequencing) [15].…”

“…The value of metagenomics to the industry is not only in the ability to produce information-rich data at low cost, but also the ability for the tool to be predictive in different environments. Metagenome pro les are in uenced by environment [14] and so an assessment needs to be made whether decisions based on the metagenome pro le of an animal in one environment will be stable across other environments. Furthermore, while greenhouse gas emissions and feed e ciency are obvious choices for the involvement of the metagenome, the value of harnessing metagenomic information can be extended if there is an association with other, perhaps less obvious, production traits.…”