Genetic analysis of disease resilience of wean-to-finish pigs under a natural disease challenge model using reaction norms

Background

Disease resilience is the ability to maintain performance across environments with different disease challenge loads (CL). A reaction norm describes the phenotypes that a genotype can produce across a range of environments and can be implemented using random regression models. The objectives of this study were to: (1) develop measures of CL using growth rate and clinical disease data recorded under a natural polymicrobial disease challenge model; and (2) quantify genetic variation in disease resilience using reaction norm models.

Methods

Different CL were derived from contemporary group effect estimates for average daily gain (ADG) and clinical disease phenotypes, including medical treatment rate (TRT), mortality rate, and subjective health scores. Resulting CL were then used as environmental covariates in reaction norm analyses of ADG and TRT in the challenge nursery and finisher, and compared using model loglikelihoods and estimates of genetic variance associated with CL. Linear and cubic spline reaction norm models were compared based on goodness-of-fit and with multi-variate analyses, for which phenotypes were separated into three traits based on low, medium, or high CL.

Results

Based on model likelihoods and estimates of genetic variance explained by the reaction norm, the best CL for ADG in the nursery was based on early ADG in the finisher, while the CL derived from clinical disease traits across the nursery and finisher was best for ADG in the finisher and for TRT in the nursery and across the nursery and finisher. With increasing CL, estimates of heritability for nursery and finisher ADG initially decreased, then increased, while estimates for TRT generally increased with CL. Genetic correlations for ADG and TRT were low between high versus low CL, but high for close CL. Linear reaction norm models fitted the data significantly better than the standard genetic model without genetic slopes, while the cubic spline model fitted the data significantly better than the linear reaction norm model for most traits. Reaction norm models also fitted the data better than multi-variate models.

Conclusions

Reaction norm models identified genotype-by-environment interactions related to disease CL. Results can be used to select more resilient animals across different levels of CL, high-performance animals at a given CL, or a combination of these.

Why breed disease-resilient livestock, and how?

BACKGROUND

Fighting and controlling epidemic and endemic diseases represents a considerable cost to livestock production. Much research is dedicated to breeding disease resilient livestock, but this is not yet a common objective in practical breeding programs. In this paper, we investigate how future breeding programs may benefit from recent research on disease resilience.

MAIN BODY

We define disease resilience in terms of its component traits resistance (R: the ability of a host animal to limit within-host pathogen load (PL)) and tolerance (T: the ability of an infected host to limit the damage caused by a given PL), and model the host's production performance as a reaction norm on PL, depending on R and T. Based on this, we derive equations for the economic values of resilience and its component traits. A case study on porcine respiratory and reproductive syndrome (PRRS) in pigs illustrates that the economic value of increasing production in infectious conditions through selection for R and T can be more than three times higher than by selection for production in disease-free conditions. Although this reaction norm model of resilience is helpful for quantifying its relationship to its component traits, its parameters are difficult and expensive to quantify. We consider the consequences of ignoring R and T in breeding programs that measure resilience as production in infectious conditions with unknown PL-particularly, the risk that the genetic correlation between R and T is unfavourable (antagonistic) and that a trade-off between them neutralizes the resilience improvement. We describe four approaches to avoid such antagonisms: (1) by producing sufficient PL records to estimate this correlation and check for antagonisms-if found, continue routine PL recording, and if not found, shift to cheaper proxies for PL; (2) by selection on quantitative trait loci (QTL) known to influence both R and T in favourable ways; (3) by rapidly modifying towards near-complete resistance or tolerance, (4) by re-defining resilience as the animal's capacity to resist (or recover from) the perturbation caused by an infection, measured as temporal deviations of production traits in within-host longitudinal data series.

CONCLUSIONS

All four alternatives offer promising options for genetic improvement of disease resilience, and most rely on technological and methodological developments and innovation in automated data generation.

A probabilistic Poisson-based model to detect PRRSV recirculation using sow production records

Porcine reproductive and respiratory syndrome (PRRS) is a viral disease associated with a decrease in the number of born alive piglets (NBA) and an increase in the number of lost piglets (NLP) per farrowing. Under practical conditions, it is critical to assess whether a farm is suffering PRRSV recirculation in the sow herd as soon as possible. The aim of this research work was to develop a new method to detect potential PRRSV recirculation in sow production farms. Sow reproductive performance records from one farm (farm T) were used to set up the method and records from ten additional farms (farms V1 to V10) were used for validation. A conditional Poisson model of NLP on NBA was proposed to fit the data. A three-step procedure was implemented to detect potential PRRSV recirculation: (i) computation of the maximum-likelihood estimates of the expected values of NBA and NLP in a PRRSV non-recirculating scenario; (ii) calculation, for each farrowing, of the p-value associated with the probability of jointly observing deviations towards decreased NBA and increased NLP. The detection of a potential PRRSV recirculation was based on (iii) the combined p-value resulting from weighing the p-values of the last N farrowings by the chi-square-inverse method. In order to gain specificity, a displacement on the expected non-recirculating NBA and NLP values was used for tuning purposes. With this approach, two PRRSV circulating periods were detected in farm T, which were confirmed with standard laboratorial diagnostic techniques. The method was subsequently validated in farms V1 to V10, where ten PRRSV-recirculating time episodes had been diagnosed. The method proposed here was able to detect the ten PRRSV recirculations using a relatively small set of contiguous farrowings, with only two mismatched weeks, one as a false negative, in farm V1, and one as a false positive, in farm V4. It is concluded that a conditional Poisson-based model of NLP on NBA can be a useful tool for routinely detecting PRRSV recirculation in sow herds.

Genetics and genomics of uniformity and resilience in livestock and aquaculture species: A review

Genetic control of residual variance offers opportunities to increase uniformity and resilience of livestock and aquaculture species. Improving uniformity and resilience of animals will improve health and welfare of animals and lead to more homogenous products. Our aims in this review were to summarize the current models and methods to study genetic control of residual variance, genetic parameters and genomic results for residual variance and discuss future research directions. Typically, the genetic coefficient of variation is high (median = 0.27; range 0-0.86) and the heritability of residual variance is low (median = 0.01; range 0-0.10). Higher heritabilities can be achieved when increasing the number of records per animal. Divergent selection experiments have supported the feasibility of selecting for high or low residual variance. Genomic studies have revealed associations in regions related to stress, including those from the heat shock protein family. Although the number of studies is growing, genetic control of residual variance is still poorly understood, but big data and genomics offer great opportunities.

Quantifying the health challenges in an Australian piggery using medication records for the definition of disease resilience1

Disease resilience is the ability to maintain performance and health, despite infection challenges in the environment. The evaluation of disease resilience requires measures of environment infection challenges, along with other environmental challenges. The overall objective of this study was to define disease resilience using pedigree, production, and medication records from an Australian herd of Large White pigs. The extent to which the infection challenges were captured by environmental descriptors based on contemporary group (CG) estimates of growth was assessed (n = 8,835). There were moderately negative linear relationships (r = -0.29, p = 0.08) between CG estimates (39 CGs) of growth and the frequency of medicated pigs (n = 812 medicated pigs). This suggests that CG estimates of growth partly capture health challenges. However, because the health challenges were not of the pathogenic nature for this herd, these environmental descriptors may not be appropriate for the evaluation of disease resilience. Subsequently, an alternative approach to select for health was provided, where health was defined as a binary outcome of medication status, fitted in a generalized linear mixed sire model. Two health-trait definitions were explored, which differed in the number of control (nonmedicated) pigs per litter. The 'reduced-control' health trait had a representative sample of littermates with available performance records, and the 'full-control' health trait included all piglets weaned per litter (i.e., performance-tested and non-performance-tested pigs). All 812 medicated pigs had performance records available. The remaining 8,023 pigs in the reduced-control and 21,352 pigs in the full-control health traits were assumed to have not been medicated (controls). Male pigs from litters with a higher number of postweaning deaths were more likely to be medicated for both health traits. Heritability was consistent for both trait definitions, at 0.06 ± 0.04 (± SE) (reduced-control) and 0.04 ± 0.03 (full-control). While results may be specific for individual herds depending on health status, these estimates align with those presented in literature for other health traits. Together, these results demonstrate that routinely collected medication records may be useful for pig breeding programs and their economic importance and genetic background should be explored further.

The effect of a porcine reproductive and respiratory syndrome outbreak on genetic parameters and reaction norms for reproductive performance in pigs1

The objective of this study was to estimate genetic parameters of antibody response and reproductive traits after exposure to porcine reproductive and respiratory syndrome virus. Blood samples were taken approximately 60 d after the outbreak. Antibody levels were quantified as the sample-to-positive ratio (S/P ratio) using a fluorescent microsphere assay. Reproductive traits included total number born (TNB), number born alive (NBA), number stillborn (NSB), number mummified (NBM), and number born dead (NBD). Mortality traits were log transformed for genetic analyses. Data were split into prior, during, and after the disease outbreak phases using visual appraisal of the estimates of farm-year-week effects for each reproductive trait. For NBA, data from all phases were combined into a reaction norm analysis with regression on estimates of farm-year-week effects for NBA. Heritability for S/P ratio was estimated at 0.17 ± 0.05. Heritability estimates for reproduction traits were all low and were lower during the outbreak for NBA but greater for mortality traits. TNB was not greatly affected during the outbreak, as many sows that farrowed during the outbreak were mated prior to the outbreak. Heritability for TNB decreased from 0.13 (prior) to 0.08 (after). Genetic correlation estimates between prior to and during the outbreak were high for TNB (0.86 ± 0.23) and NBA (0.98 ± 0.38) but lower for mortality traits: 0.65 ± 0.43, −0.42 ± 0.55, and 0.29 ± 1.39 for LNSB, LNBM, and LNBD, respectively. TNB prior to and after the outbreak had a lower genetic correlation (0.32 ± 0.33). In general, genetic correlation estimates of S/P ratio with reproductive performance during the outbreak were below 0.20 in absolute value, except for LNSB (−0.73 ± 0.29). Based on the reaction norm model, estimates of genetic correlations between the intercept and slope terms ranged from 0.24 ± 0.50 to 0.54 ± 0.35 depending on the parameterization used, indicating that selection for the intercept may result in indirect selection for steeper slopes, and thus, less resilient animals. In general, estimates of genetic correlations between farm-year-week effect classes based on the reaction norm model resembled estimates of genetic correlations from the multivariate analysis. Overall, compared to previous studies, antibody S/P ratios showed a lower heritability (0.17 ± 0.05) and low genetic correlations with reproductive performance during a porcine reproductive and respiratory syndrome outbreak, except for the LNSB.

Reaction norm analysis of pig growth using environmental descriptors based on alternative traits

Contemporary group (CG) estimates of different phenotypes have not been widely explored for pigs. The objective of this study was to extend the traits used to derive environmental descriptors of the growing pig, to include CG estimates of early growth between birth and start of feed intake test (EADG), growth during feed intake test (TADG), lifetime growth (ADG), daily feed intake (DFI), backfat (BF) and muscle depth (MD). Pedigree and performance records (n = 7,746) from a commercial Australian piggery were used to derive environmental descriptors based on CG estimates of these six traits. The CG estimates of growth traits described different aspects of the environment from the CG estimates of carcass traits (r < 0.10). These definitions of the environment then were used in reaction norm analysis of growth, to evaluate sire-by-environment interaction (Sire × E) for growth. The most appropriate reaction norm model to evaluate Sire × E for growth was dependent on the environmental descriptor used. If the trait used to derive an environmental descriptor was distinctly different from growth (e.g., BF and MD), CG as an additional random effect was required in the model. If not included, inflated common litter effect and sire intercept variance suggest there was unaccounted environmental variability. There was no significant Sire × E using any of the definitions of the environment, with estimated variance in sire slopes largest when environments were defined by BF ( σ ^ bi 2  = 97 ± 83 (g/day)² ), followed by environments defined by DFI ( σ ^ bi 2  = 39 ± 101 (g/day)² ). While there appears to be differences in ability to detect Sire × E, improved data structure is required to better assess these environmental descriptors based on alternative traits. The ideal trait, or combination of traits, used to derive environmental descriptors may be unique for individual herds. Therefore, multiple phenotypes should be further explored for the evaluation of Sire × E for growth in the pig.

Genotype by feed interaction for feed efficiency and growth performance traits in pigs

A major objective of pork producers is to reduce production cost. Feeding may account for over 75% of pork production costs. Thus, selecting pigs for feed efficiency (FE) traits is a priority in pig breeding programs. While in the Americas, pigs are typically fed high-input diets, based on corn and soybean meal (CS); in Western Europe, pigs are commonly fed diets based on wheat and barley with high amounts of added protein-rich coproducts (WB), e.g., from milling and seed-oil industries. These two feeding scenarios provided a realistic setting for investigating a specific type of genotype by environment interaction; thus, we investigated the genotype by feed interaction (GxF). In the presence of a GxF, different feed compositions should be considered when selecting for FE. This study aimed to 1) verify the presence of a GxF for FE and growth performance traits in different growth phases (starter, grower, and finisher) of 3-way crossbred growing-finishing pigs fed either a CS (547 boars and 558 gilts) or WB (567 boars and 558 gilts) diet; and 2) to assess and compare the expected responses to direct selection under the 2 diets and the expected correlated responses for one diet to indirect selection under the other diet. We found that GxF did not interfere in the ranking of genotypes under both diets for growth, protein deposition, feed intake, energy intake, or feed conversion rate. Therefore, for these traits, we recommend changing the diet of growing-finishing pigs from high-input feed (i.e., CS) to feed with less valuable ingredients, as WB, to reduce production costs and the environmental impact, regardless of which diet is used in selection. We found that GxF interfered in the ranking of genotypes and caused heterogeneity of genetic variance under both diets for lipid deposition (LD), residual energy intake (REI), and residual feed intake (RFI). Thus, selecting pigs under a diet different from the diet used for growing-finishing performance could compromise the LD in all growth phases, compromise the REI and RFI during the starter phase, and severely compromise the REI during the grower phase. In particular, when pigs are required to consume a WB diet for growing-finishing performance, pigs should be selected for FE under the same diet. Breeding pigs for FE under lower-input diets should be considered, because FE traits will become more important and lower-input diets will become more widespread in the near future.

Selection on resilience improves disease resistance and tolerance to infections

Response to infection in animals has 2 main mechanisms: resistance (ability to control pathogen burden) and tolerance (ability to maintain performance given the pathogen burden). Selection on disease resistance and tolerance to infections seems a promising avenue to increase productivity of animals in the presence of disease infections, but it is hampered by a lack of records of pathogen burden of infected animals. Selection on resilience (ability to maintain performance regardless of pathogen burden) may, therefore, be an alternative pragmatic approach, because it does not need records of pathogen burden. Therefore, the aim of this study was to assess response to selection in resistance and tolerance when selecting on resilience compared with direct selection on resistance and tolerance. Monte Carlo simulation was used combined with selection index theory to predict responses to selection. Using EBV for resilience in the absence of records for pathogen burden resulted in favorable responses in resistance and tolerance to infections, with higher responses in tolerance than in resistance. If resistance and tolerance were unfavorably correlated, lower selection responses were obtained, especially in resistance. When the genetic correlation was very unfavorable, the selection response in tolerance became negative. Results showed that lower selection responses in resistance and tolerance were obtained when the frequency of disease outbreaks was 10% rather than 50% of the contemporary groups. The efficiency of selection on EBV for resilience compared with selection on EBV for resistance and tolerance was, however, not affected by the frequency of disease outbreaks. When records on pathogen burden were available, selection responses in resistance, tolerance, and the total breeding goal were 3 to 28%, 66 to 398%, and 2 to 11% higher, respectively, than when using the EBV for resilience, showing a clear benefit of recording pathogen burden. This study shows that selection on resilience is a pragmatic way of increasing disease resistance and tolerance to infections in the absence of records on pathogen burden, but recording pathogen burden would yield higher selection responses in resistance and tolerance.

Contemporary group estimates adjusted for climatic effects provide a finer definition of the unknown environmental challenges experienced by growing pigs

Environmental descriptors derived from mean performances of contemporary groups (CGs) are assumed to capture any known and unknown environmental challenges. The objective of this paper was to obtain a finer definition of the unknown challenges, by adjusting CG estimates for the known climatic effects of monthly maximum air temperature (MaxT), minimum air temperature (MinT) and monthly rainfall (Rain). As the unknown component could include infection challenges, these refined descriptors may help to better model varying responses of sire progeny to environmental infection challenges for the definition of disease resilience. Data were recorded from 1999 to 2013 at a piggery in south-east Queensland, Australia (n = 31,230). Firstly, CG estimates of average daily gain (ADG) and backfat (BF) were adjusted for MaxT, MinT and Rain, which were fitted as splines. In the models used to derive CG estimates for ADG, MaxT and MinT were significant variables. The models that contained these significant climatic variables had CG estimates with a lower variance compared to models without significant climatic variables. Variance component estimates were similar across all models, suggesting that these significant climatic variables accounted for some known environmental variation captured in CG estimates. No climatic variables were significant in the models used to derive the CG estimates for BF. These CG estimates were used to categorize environments. There was no observable sire by environment interaction (Sire×E) for ADG when using the environmental descriptors based on CG estimates on BF. For the environmental descriptors based on CG estimates of ADG, there was significant Sire×E only when MinT was included in the model (p = .01). Therefore, this new definition of the environment, preadjusted by MinT, increased the ability to detect Sire×E. While the unknown challenges captured in refined CG estimates need verification for infection challenges, this may provide a practical approach for the genetic improvement of disease resilience.

Review: Deciphering animal robustness. A synthesis to facilitate its use in livestock breeding and management

As the environments in which livestock are reared become more variable, animal robustness becomes an increasingly valuable attribute. Consequently, there is increasing focus on managing and breeding for it. However, robustness is a difficult phenotype to properly characterise because it is a complex trait composed of multiple components, including dynamic elements such as the rates of response to, and recovery from, environmental perturbations. In this review, the following definition of robustness is used: the ability, in the face of environmental constraints, to carry on doing the various things that the animal needs to do to favour its future ability to reproduce. The different elements of this definition are discussed to provide a clearer understanding of the components of robustness. The implications for quantifying robustness are that there is no single measure of robustness but rather that it is the combination of multiple and interacting component mechanisms whose relative value is context dependent. This context encompasses both the prevailing environment and the prevailing selection pressure. One key issue for measuring robustness is to be clear on the use to which the robustness measurements will employed. If the purpose is to identify biomarkers that may be useful for molecular phenotyping or genotyping, the measurements should focus on the physiological mechanisms underlying robustness. However, if the purpose of measuring robustness is to quantify the extent to which animals can adapt to limiting conditions then the measurements should focus on the life functions, the trade-offs between them and the animal's capacity to increase resource acquisition. The time-related aspect of robustness also has important implications. Single time-point measurements are of limited value because they do not permit measurement of responses to (and recovery from) environmental perturbations. The exception being single measurements of the accumulated consequence of a good (or bad) adaptive capacity, such as productive longevity and lifetime efficiency. In contrast, repeated measurements over time have a high potential for quantification of the animal's ability to cope with environmental challenges. Thus, we should be able to quantify differences in adaptive capacity from the data that are increasingly becoming available with the deployment of automated monitoring technology on farm. The challenge for future management and breeding will be how to combine various proxy measures to obtain reliable estimates of robustness components in large populations. A key aspect for achieving this is to define phenotypes from consideration of their biological properties and not just from available measures.

Genetic variation for farrowing rate in pigs in response to change in photoperiod and ambient temperature

Seasonal infertility is often observed as anestrus and a lower conception rate resulting in a reduced farrowing rate (FR) during late summer and early autumn. This is often regarded as an effect of heat stress; however, we observed a reduction in the FR of sows even after correcting for ambient temperature in our data. Therefore, we added change in photoperiod in the analysis of FR considering its effect on sow fertility. Change in photoperiod was modeled using the cosine of the day of first insemination within a year. On an average, the FR decreased by 2% during early autumn with decreasing daily photoperiod compared with early summer with almost no change in daily photoperiod. It declined 0.2% per degree Celsius of ambient temperature above 19.2°C. This result is a step forward in disentangling the 2 environmental components responsible for seasonal infertility. Our next aim was to estimate the magnitude of genetic variation in FR in response to change in photoperiod and ambient temperature to explore opportunities for selecting pigs to have a constant FR throughout the year. We used reaction norm models to estimate additive genetic variation in response to change in photoperiod and ambient temperature. The results revealed a larger genetic variation at stressful environments when daily photoperiod decreased and ambient temperatures increased above 19.2°C compared with neutral environments. Genetic correlations between stressful environments and nonstressful environments ranged from 0.90 (±0.03) to 0.46 (±0.13) depending on the severity of the stress, indicating changes in expression of FR depending on the environment. The genetic correlation between responses of pigs to changes in photoperiod and to those in ambient temperature were positive, indicating that pigs tolerant to decreasing daily photoperiod are also tolerant to high ambient temperatures. Therefore, selection for tolerance to decreasing daily photoperiod should also increase tolerance to high ambient temperatures or vice versa.

Efficient Boar Semen Production and Genetic Contribution: The Impact of Low-Dose Artificial Insemination on Fertility

Diluting semen from high fertile breeding boars, and by that inseminating many sows, is the core business for artificial insemination (AI) companies worldwide. Knowledge about fertility results is the reason by which an AI company can lower the concentration of a dose. Efficient use of AI boars with high genetic merit by decreasing the number of sperm cells per insemination dose is important to maximize dissemination of the genetic progress made in the breeding nucleus. However, a potential decrease in fertility performance in the field should be weighed against the added value of improved genetics and, in general, is not tolerated in commercial production. This overview provides some important aspects that influence the impact of low-dose AI on fertility: (i) the importance of monitoring field fertility, (ii) the need for accurate and precise semen assessment, (iii) the parameters that are taken into account, (iv) the application of information from genetic and genomic selection and (v) the optimization when using different AI techniques. Efficient semen production, processing and insemination in combination with increasing use of genetic and genomic applications result in maximum impact of genetic trend.

Estimation of genetic parameters and breeding values across challenged environments to select for robust pigs

Robustness is an important issue in the pig production industry. Since pigs from international breeding organizations have to withstand a variety of environmental challenges, selection of pigs with the inherent ability to sustain their productivity in diverse environments may be an economically feasible approach in the livestock industry. The objective of this study was to estimate genetic parameters and breeding values across different levels of environmental challenge load. The challenge load (CL) was estimated as the reduction in reproductive performance during different weeks of a year using 925,711 farrowing records from farms distributed worldwide. A wide range of levels of challenge, from favorable to unfavorable environments, was observed among farms with high CL values being associated with confirmed situations of unfavorable environment. Genetic parameters and breeding values were estimated in high- and low-challenge environments using a bivariate analysis, as well as across increasing levels of challenge with a random regression model using Legendre polynomials. Although heritability estimates of number of pigs born alive were slightly higher in environments with extreme CL than in those with intermediate levels of CL, the heritabilities of number of piglet losses increased progressively as CL increased. Genetic correlations among environments with different levels of CL suggest that selection in environments with extremes of low or high CL would result in low response to selection. Therefore, selection programs of breeding organizations that are commonly conducted under favorable environments could have low response to selection in commercial farms that have unfavorable environmental conditions. Sows that had experienced high levels of challenge at least once during their productive life were ranked according to their EBV. The selection of pigs using EBV ignoring environmental challenges or on the basis of records from only favorable environments resulted in a sharp decline in productivity as the level of challenge increased. In contrast, selection using the random regression approach resulted in limited change in productivity with increasing levels of challenge. Hence, we demonstrate that the use of a quantitative measure of environmental CL and a random regression approach can be comprehensively combined for genetic selection of pigs with enhanced ability to maintain high productivity in harsh environments.