Breeding against infectious diseases in animals

Infectious diseases in farm animals are of major concern because of animal welfare, production costs, and public health. Farms undergo huge economic losses due to infectious disease. The costs of infections in farm animals are mainly due to production losses, treatment of infected animals, and disease control strategies. Control strategies, however, are not always successful. Selective breeding for the animals that can mount a defence against infection could therefore be a promising approach. Defensive ability of an animal has two main mechanisms: resistance (ability to control the pathogen burden) and tolerance (ability to maintain performance when pathogen burden increases). When it is difficult to distinguish between resistance and tolerance, defensive ability is measured as resilience that is the ability to maintain performance during a disease outbreak regardless of pathogen burden. Studies have focused on the genetics of resistance and resilience with little known about the genetics of tolerance and its relationship with resistance and resilience. The objectives of this thesis were to: 1) estimate the genetic variation in resistance, tolerance, and resilience to infection in order to assess the amenability of these traits for selective breeding in farm animals, 2) estimate the genetic correlation between resistance, tolerance and resilience and 3) detect genomic regions associated with resistance, tolerance, and resilience.

In chapter 2, we studied the variation among sows in response to porcine reproductive and respiratory syndrome (PRRS). First a statistical method was developed to detect PRRS outbreaks based on reproduction records of sows. The method showed a high sensitivity (78%) for disease phases. Then the variation of sows in response to PRRS was quantified using 2 models on the traits number of piglets born alive (NBA) and number of piglets born dead (LOSS): 1) bivariate model considering the trait in healthy and disease phases as different traits, and 2) reaction norm model modelling the response of sows as a linear regression of the trait on herd-year-week estimates of NBA. Trait correlations between healthy and disease phases deviated from unity (0.57±0.13 – 0.87±0.18). The repeatabilities ranged from 0.07±0.027 to 0.16±0.005. The reaction norm model had higher predictive ability in disease phase compared to the bivariate model.

In chapter 3 we studied 1) the genetic variation in resistance and tolerance of sheep to gastrointestinal nematode infection and 2) the genetic correlation between resistance and tolerance. Sire models on faecal nematode egg count (FEC), IgA, and pepsinogen were used to study the genetic variation in resistance. Heritability for resistance traits ranged from 0.19±0.10 to 0.59±0.20. A random regression model was used to study the reaction norm of sheep body weight on FEC as an estimate of tolerance to nematode infection. We observed a significant genetic variance in tolerance (P<0.05). Finally a bivariate model was used to study the genetic correlation between resistance and tolerance. We observed a negative genetic correlation (-0.63±0.25) between resistance and tolerance.

In chapter 4, we studied the response to selection in resistance and tolerance when using estimated breeding values for resilience. We used Monte Carlo simulation to generate 100 half-sib families with known breeding values for resistance (pathogen burden) and tolerance. We used selection index theory to predict response to selection for resistance and tolerance: 1) when pathogen burden is known and selection is based on true breeding values for resistance and tolerance and 2) when pathogen burden is unknown and selection is based on estimated breeding values for resilience. Using EBV for resilience in absence of records for pathogen burden resulted in favourable responses in resistance and tolerance to infections, with more emphasis on tolerance than on resistance. However, more genetic gain in resistance and tolerance could be achieved when pathogen burden was known.

In chapter 5 we studied genomics regions associated with resistance, resilience, and tolerance to PRRS. Resistance was modelled as sire effect on area under the PRRS viremia curve up to 14 days post infection (AUC14). Resilience was modelled as sire effects on daily growth of pigs up to 28 days post infection (ADG28). Tolerance was modelled as the sire effect on the regression of ADG28 on AUC14. We identified a major genomics region on chromosome 4 associated with resistance and resilience to PRRS. We also identified genomics regions on chromosome 1 associated with tolerance to PRRS.

In the general discussion (chapter 6) I discussed: 1) response to infection as a special case of genotype by environment interaction, 2) random regression model as a statistical tool for studying response to disease, 3) advantages and requirements of random regression models, and 4) selective breeding of farm animals for resistance, tolerance, and resilience to infections. I concluded that random regression is a powerful approach to estimate response to infection in animals. If the adequate amount of data is available random regression model could estimate breeding values of animals more accurately compared to other models. I also concluded that before including resistance and tolerance into breeding programs, breeders should make sure about the added values of including these traits on genetic progress. Selective breeding for resilience could be a pragmatic approach to simultaneously improve resistance and tolerance.