Breeding for robustness: investigating the genotype-by-environment interaction and micro-environmental sensitivity of Genetically Improved Farmed Tilapia (Oreochromis niloticus).

Robustness has become a highly desirable breeding goal in the globalized agricultural market. Both genotype-by-environment interaction (G × E) and micro-environmental sensitivity are important robustness components of aquaculture production, in which breeding stock is often disseminated to different environments. The objectives of this study were (i) to quantify the degree of G × E by assessing the growth performance of Genetically Improved Farmed Tilapia (GIFT) across three countries (Malaysia, India and China) and (ii) to quantify the genetic heterogeneity of environmental variance for body weight at harvest (BW) in GIFT as a measure of micro-environmental sensitivity. Selection for BW was carried out for 13 generations in Malaysia. Subsets of 60 full-sib families from Malaysia were sent to China and India after five and nine generations respectively. First, a multi-trait animal model was used to analyse the BW in different countries as different traits. The results indicate a strong G × E. Second, a genetically structured environmental variance model, implemented using Bayesian inference, was used to analyse micro-environmental sensitivity of BW in each country. The analysis revealed the presence of genetic heterogeneity of both BW and its environmental variance in all environments. The presence of genetic variation in residual variance of BW implies that the residual variance can be modified by selection. Incorporating both G × E and micro-environmental sensitivity information may help in selecting robust genotypes with high performance across environments and resilience to environmental fluctuations.

Using RNA-Seq SNP data to reveal potential causal mutations related to pig production traits and RNA editing.

RNA-Seq technology is widely used in quantitative gene expression studies and identification of non-annotated transcripts. However this technology also can be used for polymorphism detection and RNA editing in transcribed regions in an efficient and cost-effective way. This study used SNP data from an RNA-Seq assay to identify genes and mutations underlying production trait variations in an experimental pig population. The hypothalamic and hepatic transcriptomes of nine extreme animals for growth and fatness from an (Iberian × Landrace) × Landrace backcross were analyzed by RNA-Seq methodology, and SNP calling was conducted. More than 125 000 single nucleotide variants (SNVs) were identified in each tissue, and 78% were considered to be potential SNPs, those SNVs segregating in the context of this study. Potential informative SNPs were detected by considering those showing a homozygous or heterozygous genotype in one extreme group and the alternative genotype in the other group. In this way, 4396 and 1862 informative SNPs were detected in hypothalamus and liver respectively. Out of the 32 SNPs selected for validation, 25 (80%) were confirmed as actual SNPs. Association analyses for growth, fatness and premium cut yields with 19 selected SNPs were carried out, and four potential causal genes (RETSAT, COPA, RNMT and PALMD) were identified. Interestingly, new RNA editing modifications were detected and validated for the NR3C1:g.102797 (ss1985401074) and ACSM2B:g.13374 (ss1985401075) positions and for the COG3:g3.4525 (ss1985401087) modification previously identified across vertebrates, which could lead to phenotypic variation and should be further investigated.

Genetic control of the environmental variance for birth weight in seven generations of a divergent selection experiment in mice.

Data from seven generations of a divergent selection experiment designed for environmental variability of birth weight were analysed to estimate genetic parameters and to explore signs of selection response. A total of 10 783 birth weight records from 638 females and 1127 litters in combination with 10 007 pedigree records were used. Each record of birth weight was assigned to the mother of the pup in a heteroscedastic model, and after seven generations of selection, evidence of success in the selection process was shown. A Bayesian analysis showed that success of the selection process started from the first generation for birth weight and from the second generation for its environmental variability. Genetic parameters were estimated across generations. However, only from the third generation onwards were the records useful to consider the results to be reliable. The results showed a consistent positive and low genetic correlation between the birth weight trait and its environmental variability, which could allow an independent selection process. This study has demonstrated that the genetic control of the birth weight environmental variability is possible in mice. Nevertheless, before the results are applied directly in farm animals, it would be worth confirming any other implications on other important traits, such as robustness, longevity and welfare.

Transcriptional analysis of intramuscular fatty acid composition in the longissimus thoracis muscle of Iberian × Landrace back-crossed pigs.

This study aimed at identifying differential gene expression conditional on the fatty acid profile of the longissimus thoracis (Lt) muscle, a prime cut of economic relevance for fresh and cured pork production. A population of 110 Iberian (25%) × Landrace (75%) back-crossed pigs was used, because these two breeds exhibit extreme profiles of intramuscular saturated fatty acid, monounsaturated fatty acid (MUFA) and polyunsaturated fatty acid (PUFA) contents. Total RNA from Lt muscle was individually hybridized to GeneChip Porcine Genome arrays (Affymetrix). A principal component analysis was performed with data from the 110 animals to select 40 extreme animals based on the total fatty acid profile and the MUFA composition (MAP). Comparison of global transcription levels between extreme fatty acid profile pigs (n = 40) resulted in 219 differentially expressed probes (false discovery rate <0.10). Gene ontology, pathway and network analysis indicated that animals with higher percentages of PUFA exhibit a shift toward a more oxidative muscular metabolism state, with a raise in mitochondria function (PPARGC1A, ATF2), fatty acid uptake and oxidation (FABP5, MGLL). On the other hand, 87 probes were differentially expressed between MUFA composition groups (n = 40; false discovery rate <0.10). In particular, muscles rich in n-7 MUFA expressed higher levels of genes involved in lipid metabolism (GLUL, CRAT, PLA2G15) and lower levels of fatty acid elongation genes (ELOVL5). Moreover, the chromosomal position of FABP5, PAQR3, MGLL, PPARGC1A, GLUL and ELOVL5 co-localized with very relevant QTL for fat deposition and composition described in the same resource population. This study represents a complementary approach to identifying genes underlying these QTL effects.

Genetic parameters for birthweight environmental variability in mice.

Data from a divergent experiment for birthweight (BrW) environmental variability were used to estimate genetic parameters for BrW trait and its environmental variability by fitting both homoscedastic (HO) and heteroscedastic (HE) models. A total of 5 475 records of BrW from animals born from inbred dams, and 7 140 pedigree records were used. The heritability of BrW using the model HO was 0.27, with the litter effect much more important, 0.43. The model HE provided a genetic correlation between the trait and its environmental variability that was very high and negative, -0.97, and a high value for the additive genetic variance for environmental variability, suggesting an artefact in the model. The residual skewness was found to be essentially null. A model considering the genetic correlation null was also fitted, and used to obtain the breeding values for the selection process. Moreover, the trait was considered as maternal resulting in similar estimates under the model HO, but more reasonable for the genetic correlation between the trait and its environmental variability of 0.48 with a value of 0.25 for the additive genetic variance regarding environmental variability under the model HE. This led to the conclusion that environmental variability of BrW in mice must be selected via dams. Estimated parameters in a reduced dataset without inbred animals did not substantially change this conclusion.

Evaluation of the porcine ACSL4 gene as a candidate gene for meat quality traits in pigs.

Long-chain acyl-CoA synthetase (ACSL) family members catalyse the formation of long-chain acyl-CoA from fatty acid, ATP and CoA, thus playing an important role in both de novo lipid synthesis and fatty acid catabolism. Previous studies in our group evaluated ACSL4 as a positional candidate gene for quantitative trait loci located on chromosome X in an Iberian × Landrace cross. A DQ144454:c.2645G>A SNP located in the 3' untranslated region of the ACSL4 gene was associated with the percentages of oleic and monounsaturated fatty acids. The aim of the present work was to evaluate the functional implication of this genetic variant. An expression analysis was performed for 120 individuals with different genotypes for the DQ144454:c.2645G>A polymorphism using real-time quantitative PCR. Differences between genotypes were identified in liver, with the ACSL4 mRNA expression levels higher in animals with the G allele than in animals with the A allele. A SNP genome-wide association study with ACSL4 relative expression levels showed significant positions on chromosomes 6 and 12. Description of positional candidate genes for ACSL4 regulation on chromosomes 6 and 12 is provided.

Restricting inbreeding while maintaining selection response for weight gain in Mus musculus.

An experiment with mice was designed to test the relative efficiency of three selection methods that help to minimize the rate of inbreeding during selection. A common house mice (Mus musculus) population was selected for 17 generations to increase the weight gain between 21 and 42 days. The population was split at random into three lines A, B and C where three selection methods were applied: individual selection and random mating, weighted selection with random mating and individual selection with minimum coancestry mating, respectively. There were three replicates for each line. Cumulated selection response was similar in the three lines, but there were differences in the level of inbreeding attained (in percentage): 31.24 (method A), 24.72 (method B) and 27.88 (method C). As consequence, lines B and C (weighted selection and minimum coancestry) showed a lower value of deterioration of fitness traits (the intrauterine mortality and the mortality at birth) than line A (random mating).

GSEVM v.2: MCMC software to analyze genetically structured environmental variance models.

This note provides a description of software that allows to fit Bayesian genetically structured variance models using Markov chain Monte Carlo (MCMC). The gsevm v.2 program was written in Fortran 90. The DOS and Unix executable programs, the user's guide, and some example files are freely available for research purposes at http://www.bdporc.irta.es/estudis.jsp. The main feature of the program is to compute Monte Carlo estimates of marginal posterior distributions of parameters of interest. The program is quite flexible, allowing the user to fit a variety of linear models at the level of the mean and the logvariance.

The effect of divergent selection for intramuscular fat on the domestic rabbit genome.

An experiment of divergent selection for intramuscular fat was carried out at Universitat Politècnica de València. The high response of selection in intramuscular fat content, after nine generations of selection, and a multidimensional scaling analysis showed a high degree of genomic differentiation between the two divergent populations. Therefore, local genomic differences could link genomic regions, encompassing selective sweeps, to the trait used as selection criterion. In this sense, the aim of this study was to identify genomic regions related to intramuscular fat through three methods for detection of selection signatures and to generate a list of candidate genes. The methods implemented in this study were Wright's fixation index, cross population composite likelihood ratio and cross population - extended haplotype homozygosity. Genomic data came from the 9th generation of the two populations divergently selected, 237 from Low line and 240 from High line. A high single nucleotide polymorphism (SNP) density array, Affymetrix Axiom OrcunSNP Array (around 200k SNPs), was used for genotyping samples. Several genomic regions distributed along rabbit chromosomes (OCU) were identified as signatures of selection (SNPs having a value above cut-off of 1%) within each method. In contrast, 8 genomic regions, harbouring 80 SNPs (OCU1, OCU3, OCU6, OCU7, OCU16 and OCU17), were identified by at least 2 methods and none by the 3 methods. In general, our results suggest that intramuscular fat selection influenced multiple genomic regions which can be a consequence of either only selection effect or the combined effect of selection and genetic drift. In addition, 73 genes were retrieved from the 8 selection signatures. After functional and enrichment analyses, the main genes into the selection signatures linked to energy, fatty acids, carbohydrates and lipid metabolic processes were ACER2, PLIN2, DENND4C, RPS6, RRAGA (OCU1), ST8SIA6, VIM (OCU16), RORA, GANC and PLA2G4B (OCU17). This genomic scan is the first study using rabbits from a divergent selection experiment. Our results pointed out a large polygenic component of the intramuscular fat content. Besides, promising positional candidate genes would be analysed in further studies in order to bear out their contributions to this trait and their feasible implications for rabbit breeding programmes.

Genetic parameters and direct, maternal and heterosis effects on litter size in a diallel cross among three commercial varieties of Iberian pig.

The Iberian pig is one of the pig breeds that has the highest meat quality. Traditionally, producers have bred one of the available varieties, exclusively, and have not used crosses between them, which has contrasted sharply with other populations of commercial pigs for which crossbreeding has been a standard procedure. The objective of this study was to perform an experiment under full diallel design among three contemporary commercial varieties of Iberian pig and estimate the additive genetic variation and the crossbreeding effects (direct, maternal and heterosis) for prolificacy. The data set comprised 18 193 records for total number born and number born alive from 3800 sows of three varieties of the Iberian breed (Retinto, Torbiscal and Entrepelado) and their reciprocal crosses (Retinto × Torbiscal, Torbiscal × Retinto, Retinto × Entrepelado, Entrepelado × Retinto, Torbiscal × Entrepelado and Entrepelado × Torbiscal), and a pedigree of 4609 individuals. The analysis was based on a multiple population repeatability model, and we developed a model comparison test that indicated the presence of direct line, maternal and heterosis effects. The results indicated the superiorities of the direct line effect of the Retinto and the maternal effect of the Entrepelado populations. All of the potential crosses produced significant heterosis, and additive genetic variation was higher in the Entrepelado than it was in the other two populations. The recommended cross for the highest yield in prolificacy is a Retinto father and an Entrepelado mother to generate a hybrid commercial sow.

Five genomic regions have a major impact on fat composition in Iberian pigs.

The adipogenic nature of the Iberian pig defines many quality attributes of its fresh meat and dry-cured products. The distinct varieties of Iberian pig exhibit great variability in the genetic parameters for fat deposition and composition in muscle. The aim of this work is to identify common and distinct genomic regions related to fatty acid composition in Retinto, Torbiscal, and Entrepelado Iberian varieties and their reciprocal crosses through a diallelic experiment. In this study, we performed GWAS using a high density SNP array on 382 pigs with the multimarker regression Bayes B method implemented in GenSel. A number of genomic regions showed strong associations with the percentage of saturated and unsaturated fatty acid in intramuscular fat. In particular, five regions with Bayes Factor >100 (SSC2 and SSC7) or >50 (SSC2 and SSC12) explained an important fraction of the genetic variance for miristic, palmitoleic, monounsaturated (>14%), oleic (>10%) and polyunsaturated (>5%) fatty acids. Six genes (RXRB, PSMB8, CHGA, ACACA, PLIN4, PLIN5) located in these regions have been investigated in relation to intramuscular composition variability in Iberian pigs, with two SNPs at the RXRB gene giving the most consistent results on oleic and monounsaturated fatty acid content.

Skew distribution of founder-specific inbreeding depression effects on the longevity of Landrace sows.

Inbreeding is a biological phenomenon of special relevance in domestic species in which its influence has been typically associated with reductions in animal fitness. Nevertheless, recent research suggests substantial founder-specific variability in terms of inbreeding depression on some productive traits, although this centred on a very reduced number of founders. This research focuses on the modelling of founder-specific inbreeding depression (FSID) effects from a more general point of view, characterizing the expected distribution of FSID effects on sow longevity. Under a change-point Weibull proportional hazards model solved through Bayesian inference, a skew-normal a priori distribution for the FSID effects of 19 founders was assumed. In terms of the deviance information criterion, this model was clearly preferred to other prior distributions for FSID effects as well as to a standard analysis of the overall inbreeding depression effect, although all models were consistent with an overall negative genetic effect of inbreeding on sow longevity. The joint posterior distribution of FSID effects showed a skewed pattern with substantial right-tail overexpression, in which the mean (0.036), mode (0.034), S.D. (0.032) and asymmetry parameter (0.045) reported a higher incidence of positive estimates (85.2%) with an unfavourable effect on sow longevity. The founder with the worst inbreeding depression effect reduced sow longevity by 32 days for 1% or 167 days for 10% partial inbreeding. As a whole, our analyses highlighted substantial variability in FSID effects, with unfavourable, neutral and even favourable influences on sow longevity. This heterogeneity could be related to an uneven distribution of the recessive deleterious genetic load among founder genomes, and also with the different selection pressures applied to each founder line. The implementation of skew-normal priors also proved an appealing way to bypass the strict scenario imposed by the standard symmetric-Gaussian distribution, allowing right- and left-tail overexpression as well as non-zero modal estimates.

Bayesian analysis of quantitative traits using skewed distributions.

Statistical models for genetic evaluation often make use of Gaussian distributions. However, some new statistical developments allow the use of an asymmetric distribution for the residuals. Within this context, we analysed three different patterns for the residual term on a data set consisting of 63 208 litter-size records, belonging to 19 255 sows, with a pedigree including 27 911 individuals. The three different residual distributions were: (1) Gaussian distribution, (2) asymmetric Gaussian distribution and (3) asymmetric Gaussian distribution with a hierarchical scheme for the asymmetry parameter. The operational model always included order of parity and herd-year-season as systematic effects, and the permanent environmental and infinitesimal genetic effect of each sow as random effects. The most suitable model using the deviance information criterion (DIC) and posterior predictive checking was model 3. This implies systematic, additive genetic and permanent environmental control of both litter size and the asymmetry parameter of the residual distribution. The asymmetry parameter can be understood as a measure of sensitivity to negative (or positive) environmental influences on phenotypes. The posterior mean (standard deviation) of the additive genetic variance was 0.28 (0.06) for litter size and 0.07 (0.01) for the asymmetry parameter. The posterior mean (standard deviation) of the additive genetic correlation between litter size and the asymmetry parameter was 0.21 (0.07).

Modulating birth weight heritability in mice.

Expected genetic response is proportional to the heritability of the trait, and this parameter is considered inherent of a specific trait in a particular population. However, models assuming heterogeneity in residual variance lead to different estimates of heritability across combinations of systematic (environmental) effects. Modifying the residual variance of the birth weight by artificial selection was shown to be feasible in a divergent selection experiment in mice. The objectives of this work were to 1) estimate the evolution of the heritability of birth weight in mice in the mentioned experiment, and 2) estimate different heritability regarding systematic effects. Data came from eleven generations of a divergent selection experiment to modify the residual variability of birth weight in mice. A total of 15,431 birth weight records from 959 females and 1,641 litters in combination with 14,786 pedigree records were used. The model used for analysis included generation, litter size, sex, and parity number as systematic effects. Each record of birth weight was assigned to the mother of the pup in the model which assumes that the residual variance is heterogeneous and partially under genetic control. Differences in heritability between lines reached values of 0.06 in the last generations. Choosing the most extreme values of systematic effects, the birth weight heritability ranged from 0.04 to 0.22. From these results, the possibility of modulating the heritability for this trait could be explored in 1 of 2 ways: selecting to decrease the residual variability, or choosing the specific levels of the systematic effects.

Correlated genetic trends for production and welfare traits in a mouse population divergently selected for birth weight environmental variability.

The objective of this work was to study the changes that, selecting for environmental variability of birth weight (BW), could bring to other interesting traits in livestock such as: survivability at weaning (SW), litter size (LS) and weaning weight (WW), their variability assessed from standard deviations of LS, standard deviation of WW (SDWW) and also the total litter weight at birth (TLBW) and total litter weight at weaning. Data were registered after eight generations of a divergent selection experiment for BW environmental variability in mice. Genetic parameters and phenotypic and genetic evolution were assessed using linear homoscedastic and heteroscedastic models in which the traits were attributed to the female, except BW and WW that were in some models also attributed to the pup. Genetic correlation between the trait and variability levels was -0.81 for LS and -0.33 for WW. Clear divergent phenotypic trends were observed between lines for LS, WW and SDWW. Although animals were heavier in the high line, TLBW and at weaning was greater in the low line. Despite the negative genetic correlation that was obtained, SDWW was also higher in the high line. Heritabilities were 0.21 and 0.06, respectively, for LS and SW. Both phenotypic and genetic trends showed clear superiority of the low line over the high line for these traits, but inferior for WW. Heteroscedastic model performed similar to the homoscedastic model when there was enough information. Considering LS and survival, the low line was preferred from a welfare point of view, but its superiority from the productivity perspective was not clear. Robustness seemed higher as shown by a low variation and having a benefit to the animal welfare, but this still remains unclear. It was concluded that low variation benefits the welfare of animals.

Genetic parameters and crossbreeding effects of fat deposition and fatty acid profiles in Iberian pig lines.

The aim of this study was to estimate the genetic and environmental parameters and crossbreeding effects on fatty acid and fat traits in the Iberian pig. Our final goal is to explore target selection traits and define crossbreeding strategies. The phenotypes were obtained under intensive management from 470 animals in a diallelic experiment involving Retinto, Torbiscal, and Entrepelado lines. The data set was composed of backfat thickness at the fourth rib (BFT), intramuscular fat (IMF) in the longissimus thoracis (LT), and the fatty acid profile for IMF and subcutaneous fat (SCF) traits. Data were analyzed through a Bayesian bivariate animal model by using a reparameterization of Dickerson's model. The results obtained showed an important genetic determinism for all traits analyzed with heritability ranging from 0.09 to 0.67. The common environment litter effect also had an important effect on IMF (0.34) and its fatty acid composition (0.06-0.53) at slaughter. The additive genetic correlation between BFT and IMF (additive genetic correlation [] = 0.31) suggested that it would be possible to improve lean growth independent of the IMF with an appropriate selection index. Furthermore, the high additive genetic correlation ( = 0.68) found between MUFA tissues would seem to indicate that either the LT or SCF could be used as the reference tissue for MUFA selection. The relevance of the crossbreeding parameters varied according to the traits analyzed. Backfat thickness at the fourth rib and the fatty acid profile of the IMF showed relevant differences between crosses, mostly due to line additive genetic effects associated with the Retinto line. On the contrary, those for IMF crosses were probably mainly attributable to heterosis effects. Particularly, heterosis effects were relevant for the Retinto and Entrepelado crosses (approximately 16% of the trait), which could be valuable for a crossbreeding system involving these lines.

Crossbreeding effects on pig growth and carcass traits from two Iberian strains.

An experiment of a 2×2 full diallelic cross between two contemporary Iberian pig strains (Retinto: RR, and Torbiscal: TT) was conducted to estimate the crossbreeding effects for growth and carcass traits. Phenotypic records were obtained under intensive management and consisted of two different data sets. The first set comprised growth traits until weaning and was collected at two different farms (6236 and 1208 records, respectively). Specific data included individual piglet weight at birth and at weaning at 28 days and average daily gain from birth to weaning at 28 days (ADG28) for both RR and TT and their reciprocal crosses. The second set comprised growth data from birth to slaughter (~340 days and ~160 kg) and carcass traits from 349 individuals (randomly) sampled at weaning from the first dataset. Data were analyzed through a Bayesian analysis by using a reparameterization of Dickerson's model that allowed estimation of the posterior distributions of the following crossbreeding effects: average maternal breed effect (gM), average paternal breed effect (gP) and individual heterosis (hI). Results showed that the relative magnitude of crossbreeding effects depends on the trait analyzed. Crosses where Torbiscal strain was used as mother (RT and TT) achieved the greatest performance for all growth traits at weaning, leading to remarkable gM effects. The most outstanding example is the case of ADG28 where the probability of relevance was one. In contrast, TR cross showed the greatest differences from RR cross for all growth at slaughter and carcass traits. These differences were mainly due to hI and gP crossbreeding parameters. In particular, the posterior mean of hI was more noticeable for live weight at slaughter, average daily gain at slaughter and carcass length, while gP was more relevant for hams (kg) and loins (kg) representing from 3% to10% of average performance of traits. Hence, growth traits at weaning did not reveal any notable advantage of the crossbreeding scheme because of the superiority of the Torbiscal strain with respect to its mothering ability and the small hI. However, results from growth and carcass traits at slaughter would support the implementation of a TR crossbred system. It would allow exploitation of both the gP of the Torbiscal strain and the hI between these two Iberian pig strains. Additionally, gP estimates and phenotypic differences between reciprocal crosses might suggest signs of the presence of paternal genetic imprinting in primal cuts traits.

Canalization analysis of birth weight in Bruna dels Pirineus beef cattle.

Although heteroskedasticity has been a main topic of interest in beef cattle during recent decades, the current availability of canalization models provided new insights for animal breeding programs. Within this context, birth BW (BWT) was analyzed in the Bruna dels Pirineus beef cattle breed by implementing canalization models that accounted for heterogeneous residual variances due to systematic, permanent environmental effects and additive genetic effects. Analyses were performed on BWT data from 8,130 calves born in 12 commercial breeding herds contributing to the yield recording scheme of the Bruna dels Pirineus breed. Analytical models accounted for direct additive genetic, permanent environmental, and 4 systematic effects (i.e., age of the dam, sex of the calf, birth type, and herd-year-season), and the same effects were evaluated as potential sources of variation in the residual term. Their relevance was checked by the deviance information criterion (DIC), and only residual additive genetic, permanent environmental, birth type, and herd-year-season remained in the operational model, all of them originating relevant reductions in the DIC parameter. Bruna dels Pirineus calves showed a moderate heritability of 0.30 (95% high posterior density, 0.19 to 0.40) for BWT; additional additive genetic variability was revealed in the residual term, this being positively correlated with the direct additive genetic component (0.44; 95% high posterior density, 0.37 to 0.54). Genetic trends were evaluated on both sources of additive genetic variance, and relevant patterns were identified in several herds. Although this breed did not evidence a homogeneous genetic trend for the whole population, herd-specific positive and negative trends were revealed, suggesting the plausibility of genetic selection for canalization on BWT in beef cattle breeds. These results must be viewed as a contribution to the canalization research field, providing relevant information for the breeding scheme of the Bruna dels Pirineus breed, as well as important insights about the genetic background of BWT for the beef industry worldwide.

Variability-specific differential gene expression across reproductive stages in sows.

Differential gene expression analyses typically focus on departures across mathematical expectations (i.e. mean) from two or more groups of microarrays, without considering alternative patterns of departure. Nevertheless, recent studies in humans and great apes have suggested that differential gene expression could also be characterized in terms of heterogeneous dispersion patterns. This must be viewed as a very interesting genetic phenomenon clearly linked to the regulation mechanisms of gene transcription. Unfortunately, we completely lack information about the incidence and relevance of dispersion-specific differential gene expression in livestock species, although a specific Bayes factor (BF) for testing this kind of differential gene expression (i.e. within-probe heteroskedasticity) has been recently developed. Within this context, our main objective was to characterize the incidence of dispersion-specific differential gene expression in pigs and, if possible, providing the first evidence of this phenomenon in a livestock species. We evaluated dispersion-specific differential gene expression on ovary, uterus and hypophysis samples from 22 F2 Iberian × Meishan sows, where a total of 15,252 probes were interrogated. For each tissue, heteroskedasticity of probe-specific residual variances was evaluated by three pairwise comparisons involving three physiological stages, that is, heat, 15 days of pregnancy and 45 days of pregnancy. Between 2.9% and 37.4% of the analyzed probes provided statistical evidence of within-tissue across-physiological stages dispersion-specific differential gene expression (BF >1), and between 0.1% and 3.0% of them reported decisive evidence (BF >100). It is important to highlight that <8% of the heteroskedastic probes were also linked to differential gene expression in terms of departures among the probe-specific mathematical expectation of each physiological stage. This discarded the disturbance of scale effects in a high percentage of probes and suggested that probe-specific heteroskedasticity must be viewed as an independent phenomenon within the context of differential gene expression. As a whole, our results report a remarkable incidence of dispersion-specific differential gene expression across the whole genome of the pig, establishing a very interesting starting point for further studies focused on deciphering the genetic mechanisms underlying heteroskedasticity.