A genome-wide association study of fetal response to type 2 porcine reproductive and respiratory syndrome virus challenge

Abstract

Control of porcine reproductive and respiratory syndrome (PRRS) is economically important for the swine industry worldwide. As current PRRS vaccines do not completely protect against heterologous challenge, alternative means of control, including enhanced genetic resilience, are needed. For reproductive PRRS, the genetic basis of fetal response to PRRS virus (PRRSV) infection is poorly understood. Genome-wide association studies (GWAS) were done here using data from 928 fetuses from pregnant gilts experimentally challenged with type 2 PRRSV. Fetuses were assessed for viral load in thymus (VLT), viral load in endometrium (VLE), fetal death (FD) and fetal viability (FV), and genotyped at a medium density. Collectively, 21 candidate genomic regions were found associated with these traits, seven of which overlap with previously reported QTLs for pig health and reproduction. A comparison with ongoing and related transcriptomic analyses of fetal response to PRRSV infection found differentially expressed genes within 18 candidate regions. Some of these genes have immune system functions, and have been reported to contribute to host response to PRRSV infection. The results provide new evidence about the genetic basis of fetal response to PRRSV challenge, and may ultimately lead to alternative control strategies to reduce the impact of reproductive PRRS.

Figure 1. Population structure of the fetuses.

The genetic distance was calculated using the genotype of 45,255 SNPs with Hamming distance. The plot was built with 2-D multi-dimension scaling (MDS), and presents the top 2 dimensions (C1 and C2) of the population structure. Colors of the points represent fetal preservation status: green for viable (VIA), red for meconium-stained (MEC), and black for decomposed (DEC). No substantial stratification was evident.

Figure 2. Manhattan Plot for viral load in fetal thymus (VLT), viral load in endometrium (VLE), fetal death (FD) and fetal viability (FV).

The association analysis was conducted with the least absolute shrinkage and selection operator (LASSO). A generalized linear model was used. The Y-axes shows the percentage of total genetic variance (calculated as the variance of GEBV calculated using 45,255 SNPs) that can be explained by each single SNP. Thresholds corresponding to different empirical P-values were calculated with a 1,000-run permutation analysis.

Figure 3. Least-square (LS) means of viral load in fetal thymus (VLT) for individuals with different genotypes for each of three SNPs showing significant (P < 0.05) or suggestive association (P < 0.1) with VLT.

The error bars represent the 95% confidence interval. The numbers above the X-axis represent the total number of animals with that genotype.

Figure 4. Distribution of viral load in fetal thymus (VLT) grouped by number of favoured alleles present in live fetuses.

For each of the three SNPs showing significant (P < 0.05) or suggestive (P < 0.1) association with VLT, a favoured allele was determined. For each individual fetus, the total number of favoured alleles across the three SNPs was determined. The Y-axis represents the proportion of the fetuses (number of fetuses in that VL window divided by the number of all the fetuses in the group); X-axis represents PRRSV RNA concentration (logarithm 10 target copies per mg). The distribution shows the trend that a lower proportion of individuals had high VLT when more favoured alleles were present.

Figure 5. Distribution of fetal preservation status grouped by number of favoured alleles.

For each of the SNPs significantly (P < 0.05) or suggestively (P < 0.1) associated with fetal death (FD), a favoured allele was determined. For each individual fetus, the total number of favoured alleles was determined, and the results displayed for set of SNPs based on their level of significance in the asociation analysis: a) 7 SNPs with P < 0.01, b) 15 SNPs with P-values < 0.05 or < 0.01, c) 18 SNPs with P-values < 0.1, < 0.05 or < 0.01. In each plot (A, B, C, n = 928), the area of each bar is proportional to the number of fetuses, with preservation status represeted by colour (green for viable [VIA], red for meconium-stained [MEC] and black for decomposed/dead [DEC]). The distribution shows the trend that the proportion of dead fetuses (DEC, black rectangles) decreases when more favoured alleles are present. However, the same trend was evident regardless of the number of SNPs included in the analyses (i.e. plots A, B, C look similar), and the number of favoured alleles did not affect the proportion of MEC fetuses (red rectangles), which remained relatively constant.

Figure 6. Distribution of viral load in fetal thymus (VLT) and endometrium (VLE) of live fetuses.

Concentration of NVSL 97-7985 PRRSV RNA (log 10) per mg tissue measured by in-house quantitative real-time PCR.