Genotype nutrition interactions shape how ruminants eat and perform

Why genotype rarely features in ruminant feeding advice#

The paper opens by pointing out an imbalance in animal nutrition research. For pigs and poultry, genetic description is built into standard feeding schemes, but for ruminants it barely features in frameworks such as the CSIRO feeding standards or the AFRC requirements document. Oldham argues this gap matters more now that breeding tools are becoming more powerful and more widely used, because predictions of how an animal will respond to a given feed need to account for the genetic targets that animal is working toward, without regard only to its current size or stage of life.

To make this point concrete, the paper sets out a general framework linking an animal's genotype and physiological state to its feed intake and performance. The animal is described as trying to eat enough of a properly balanced feed to meet performance targets set by its genetics and current state, covering body composition, reproduction and health. Actual intake then depends on whether feed and environment allow that desired intake to be reached, or whether physical, toxic or thermal constraints get in the way. This framing underpins the rest of the paper's discussion of diet selection and dairy cow performance, a theme explored further in other papers collected in the RAAN Conference Proceedings.

Testing whether animals choose diets to hit a genetic target#

Before turning to ruminants, the paper reviews evidence from pigs that animals actively select diets to reach genetically set growth targets. Kyriazakis and Emmans allowed young pigs to become fat or thin on single feeds of differing protein to energy ratio, then gave them free choice between a low and a high protein to energy diet from 16 to 33 kg liveweight. The fatter pigs chose diets that made them leaner and the leaner pigs chose diets that added fat, so that by 33 kg pigs of the same sex converged on the same lipid to protein ratio in the carcase, a ratio that differed between sexes. Kyriazakis and colleagues also found that boars grown from 44 to 103 kg liveweight chose progressively lower protein concentrations as they matured, and that a slower growing, more adipose genotype, Meishan pigs, selected quite differently from genetically improved Large White by Landrace pigs at a given weight.

The paper then asks whether the same pattern holds in ruminants. Growing sheep offered pairs of digestible feeds differing in protein to energy ratio, as reported by Hou and colleagues, made structured choices between them. Kyriazakis and Oldham found that sheep offered highly digestible feed choices selected diets that let them grow protein at a high rate while avoiding excess nitrogen intake, with some indication that sheep preferred protein sources likely to break down more slowly in the rumen over a highly rumen-degradable source such as urea.

Rumen balance and disease also steer feed choice#

The paper extends diet-selection work to look at what happens when feeds are less digestible and when the rumen environment itself becomes the limiting factor. Cooper and colleagues offered pregnant and non-pregnant sheep pairs of feeds of differing nitrogen content at both high and low energy density. With the more digestible feed pair, pregnant and non-pregnant sheep made broadly similar choices, which the paper reads as selection matched to metabolic demand. With the less dense feed pair this pattern broke down, and the paper interprets the choices instead as an attempt to balance degradable nitrogen against fermentable carbohydrate in the rumen, the ERDP to FME ratio used in the AFRC system. Separately, sheep offered feeds differing in estimated fermentability held back their intake once one feed broke down unusually quickly in the rumen, though the paper notes the underlying mechanism, whether rumen pH, osmolality or another factor, had not been pinned down.

Disease state also altered selection. Sheep carrying a subclinical burden of the gut worm Trichostrongylus colubriformis shifted their choice between feeds of differing protein to energy ratio so as to keep protein intake steady during the drop in appetite that the infection caused, as reported by Kyriazakis and colleagues. Whether this reflects a regulated response to the extra nitrogen losses parasitism causes, or a more general effect of immune activation, was not resolved in the paper. Either way, the authors treat it as further evidence that an animal's state, aside from its feed, has to be built into any scheme meant to predict intake and performance.

Breed differences underlying genotype nutrition interactions in dairy cows#

Turning to between-breed comparisons, the paper cites work by Oldenbroek showing genotype by nutrition interactions between Jersey cattle and other dairy breeds. The gap between breeds showed up in both how much cows ate and how much milk they produced, and that gap was narrower when cows were fed mostly roughage than when they were fed mostly concentrate. Much of this difference was attributed to a higher relative feed intake capacity in Jerseys, a pattern also noted in earlier work by Gibson and by Brigstocke and colleagues. The paper notes that scaling liveweight to the power 0.58 rather than the conventional 0.75 removed the breed difference seen under high concentrate feeding in the Oldenbroek data, but the difference persisted under high roughage feeding even with that adjustment, which the paper takes as evidence the difference is real rather than a scaling artefact.

Orskov and colleagues are cited for a related point: cows within the same breed can show consistent differences in how fast fibrous feed particles pass out of the rumen, something that could affect digestibility and roughage intake. Together these findings support the paper's broader argument that genotype by nutrition interactions are not confined to differences between distinct breeds, but can also appear as variation within a breed in traits that matter most when feed quality is lower.

The Langhill selection lines under two feeding systems#

The paper's central case study comes from the Langhill Dairy Cattle Research Centre, where Veerkamp, Simm and Oldham tracked Holstein-Friesian cows in two genetic lines across three lactations. The selected line was bred by artificial insemination to bulls with the highest available predicted transmitting ability for kilograms of fat plus protein; the control line was bred to bulls averaging zero PTA for that trait. Half of each line was managed on a high concentrate system, with a ratio of concentrate to brewer's grains to grass silage of 45:5:50 on a dry matter basis, and half on a lower concentrate system at 20:5:75. Performance was measured over the first 26 weeks of lactation, with data combined across the years 1988/89 to 1992/93.

Selected line cows produced more milk and milk solids than control line cows within each feeding system, with the gap in fat plus protein yield slightly larger on the high concentrate system. The most telling comparison in the paper is between control line cows on the high concentrate system and selected line cows on the lower concentrate system: their combined fat plus protein yields came out essentially the same, although the balance of fat to protein differed. In other words, a given level of milk solids output could be reached either with averagely bred cows fed generously or with highly bred cows fed more modestly, a result the paper reads as a genuine interaction between breeding and feeding system rather than a simple additive effect of the two.

Rising genetic merit, feed intake and the metabolic gap#

Regression analysis of the Langhill data, reported in a further paper by Veerkamp and colleagues, found that on the high concentrate system increasing PTA for fat plus protein came with sizeable, statistically meaningful increases in milk solids yield and in dry matter intake alike. On the lower concentrate system, milk solids yield still rose with PTA, though at a slightly lower rate, but the associated rise in dry matter intake was small and not statistically significant. The paper interprets this as showing that, with the more digestible high concentrate diet, cows with higher genetic merit were able to raise their intake enough to meet rising demand, whereas on the less digestible diet the feed itself limited how far intake could climb.

Even under the high concentrate system, the paper notes that the increase in metabolisable energy demand associated with higher PTA exceeded the increase in energy intake, leaving a shortfall that worked out close to 0.5 MJ of extra ME needed per day for each unit rise in PTA for fat plus protein, a figure that was similar across both feeding systems. The paper reads this gap as evidence that selection for milk solids yield may be, in effect, also selecting for a greater capacity to mobilise body tissue. It raises the question of whether that capacity can keep expanding across further generations while cows stay in good condition, and suggests that breeding evaluations conducted mainly under generous, high quality feeding may not expose variation in traits that matter for performance on forage-based systems, referencing a re-ranking of sires between countries seen in the CAN2 trial.

Why predicting feed intake is only half the problem#

Having laid out evidence for genotype-driven diet selection and genotype by feeding system interactions, the paper turns to what it calls the other main unsolved problem in ruminant nutrition: once an animal's intake is known, rules are still needed to describe how the absorbed nutrients are partitioned between competing uses such as growth, milk, reproduction and maintenance. The paper notes that most ruminants in practice are routinely undernourished, malnourished, or both, relative to their genetic targets, so partition rules under constrained conditions are the practically important case, not the exception.

The paper acknowledges that biochemical simulation models, such as those reviewed by Baldwin and colleagues, attempt to describe nutrient use at a detailed metabolic level, but argues that convincing evidence this is the right level for predicting whole-animal responses is still lacking. This is presented as an open challenge rather than a settled area, and the paper does not claim to resolve it. Given that the diet-selection evidence comes mainly from relatively small, controlled comparisons in sheep and pigs, we would treat the general case for building genotype and behaviour into ruminant feeding systems as a reasoned argument worth testing further, rather than as a fully proven set of rules ready for direct use on commercial farms. The broader conference record it belongs to is held within the Recent Advances in Animal Nutrition collection, part of the wider Livestock Library index of Australian livestock research.

Sources and further reading#

• Trove library search: find a library that holds the paper
• Meat & Livestock Australia: red meat industry research and marketing body
• Sheep Genetics (LAMBPLAN and MERINOSELECT): which runs LAMBPLAN and MERINOSELECT
• BREEDPLAN genetic evaluation: genetic evaluation system for Australian beef cattle

Questions#

— The Livestock Library team