Linkage disequilibrium (LD) across the genome is critical information for association studies and genomic selection because it determines the number of SNP that should be used for a successful association analysis and genomic selection. Linkage disequilibrium also influences the accuracy of genomic breeding values. Some studies have demonstrated that SNP in strong LD are organized into discrete blocks of haplotypes, which are separated by possibly hot spots of recombination. To reduce the number of markers needed to be genotyped for association mapping, a set of SNP can be selected that labels all haplotype blocks. We estimated the LD, calculated the average haplotype block size for 6 pig lines, and compared the block size between lines. Six commercial pig lines were genotyped using the Illumina PorcineSNP60 (number of markers M = 62,163) Genotyping BeadChip (Illumina Inc.); on average, a panel of 37,623 SNP with an average minor allelic frequency (MAF) of 0.283 was included in the analysis. The LD declined as a function of distance. All pig lines had an average r(2) above 0.3 for markers 100 to 150 apart. The estimated average block size was 394.885 kb, and blocks between 100 and 400 kb were most prominent (49.96%) in all lines. These results showed that the extent of LD in pigs is much larger than in the cattle population, in accordance with the genetic map length of pigs, which is much shorter than cattle. The evaluated lines have 2,640 to 3,037 blocks, covering 45% of the pig genome, on average. Differences in haplotype block size between lines were observed for some chromosomes (i.e., SSC 3, 5, 7, 13, 14, and 18), which provide a direction for future studies of haplotype block conservation or divergence across lines.

Seventy-two pigs of three genetic groups (Brazilian indigenous breed Piau, Commercial line and Crossbred) of both sexes were slaughtered at four live weights (30, 60, 90 and 120 kg). Intramuscular fat (IMF) content in Longissimus dorsi muscle of each animal was extracted and correlated with candidate gene mRNA expression (ATN1, EEF1A2, FABP3, LDLR, MGP, OBSCN, PDHB, TRDN and RYR1). Within slaughter weight of 120 kg, Piau and Crossbred pigs showed higher IMF content (p < 0.05) than commercial animals, with 2.48, 2.08 and 1.00% respectively. Barrows presented higher values of IMF (p < 0.05) than gilts (1.54 and 1.30% respectively). Gene expression of EEF1A2, FABP3, LDLR, OBSCN, PDHB, TRDN and RYR1 were correlated with IMF (p < 0.05) using the whole dataset. For Piau data only, expression of FABP3, LDLR, MGP, OBSCN, PDHB, TRDN and RYR1 showed correlation with IMF (p < 0.05). Genes that have important roles in lipid transportation inside the cell (FABP3) and tissues (LDLR) showed correlation with IMF of, respectively, 0.68 and 0.63 using the whole data set, and 0.90 and 0.91 using data from Piau animals. The highly positive correlation of the LDLR and FAPB3 expression with IMF content may confirm that these genes are important for fat deposition in the porcine L. dorsi muscle.

BackgroundGenomic selection and genomic wide association studies are widely used methods that aim to exploit the linkage disequilibrium (LD) between markers and quantitative trait loci (QTL). Securing a sufficiently large set of genotypes and phenotypes can be a limiting factor that may be overcome by combining data from multiple breeds or using crossbred information. However, the estimated effect of a marker in one breed or a crossbred can only be useful for the selection of animals in another breed if there is a correspondence of the phase between the marker and the QTL across breeds. Using data of five pure pig (Sus scrofa) lines (SL1, SL2, SL3, DL1, DL2), one F1 cross (DLF1) and two commercial finishing crosses (TER1 and TER2), the objectives of this study were: (i) to compare the equality of LD decay curves of different pig populations; and (ii) to evaluate the persistence of the LD phase across lines or final crosses.ResultsAlmost all of the lines presented different extents of LD, except for the SL2 and DL3, both of which exhibited the same extent of LD. Similar levels of LD over large distances were found in crossbred and pure lines. The crossbred animals (DLF1, TER1 and TER2) presented a high persistence of phase with their parental lines, suggesting that the available porcine single nucleotide polymorphism (SNP) chip should be dense enough to include markers that have the same LD phase with QTL across crossbred and parental pure lines. The persistence of phase across pure lines varied considerably between the different line comparisons; however, correlations were above 0.8 for all line comparisons when marker distances were smaller than 50 kb.ConclusionsThis study showed that crossbred populations could be very useful as a reference for the selection of pure lines by means of the available SNP chip panel. Here, we also pinpoint pure lines that could be combined in a multiline training population. However, if multiline reference populations are used for genomic selection, the required density of SNP panels should be higher compared with a single breed reference population.Electronic supplementary materialThe online version of this article (doi:10.1186/s12863-014-0126-3) contains supplementary material, which is available to authorized users.

This study aimed to compare two different Genome-Wide Selection (GWS) methods (Ridge Regression BLUP À RR-BLUP and Bayesian LASSO À BL) to predict the genomic estimated breeding values (GEBV) of four phenotypes, including two boar taint compounds, i.e., the concentrations of androstenone (andro) and skatole (ska), and two carcass traits, i.e., backfat thickness (fat) and loin depth (loin), which were measured in a commercial male pig line. Six hundred twenty-two boars were genotyped for 2,500 previously selected single nucleotide polymorphisms (SNPs). The accuracies of the GEBV using both methods were estimated based on Jack-knife cross-validation. The BL showed the best performance for the andro, ska and loin traits, which had accuracy values of 0.65, 0.58 and 0.33, respectively; for the fat trait, the RR-BLUP accuracy of 0.61 outperformed the BL accuracy of 0.56. Considering that BL was more accurate for the majority of the traits, this method is the most favoured for GWS under the conditions of this study. The most relevant SNPs for each trait were located in the chromosome regions that were previously indicated as QTL regions in other studies, i.e., SSC6 for andro and ska, SSC2 for fat, and SSC11, SSC15 and SSC17 for loin.

Genome-wide association studies (GWASes) have been performed to search for genomic regions associated with residual feed intake (RFI); however inconsistent results have been obtained. A meta-analysis may improve these results by decreasing the false-positive rate. Additionally, pathway analysis is a powerful tool that complements GWASes, as it enables identification of gene sets involved in the same pathway that explain the studied phenotype. Because there are no reports on GWAS pathways-based meta-analyses for RFI in beef cattle, we used several GWAS results to search for significant pathways that may explain the genetic mechanism underlying this trait. We used an efficient permutation hypothesis test that takes into account the linkage disequilibrium patterns between SNPs and the functional feasibility of the identified genes over the whole genome. One significant pathway (valine, leucine and isoleucine degradation) related to RFI was found. The three genes in this pathway-methylcrotonoyl-CoA carboxylase 1 (MCCC1), aldehyde oxidase 1 (AOX1) and propionyl-CoA carboxylase alpha subunit (PCCA)-were found in three different studies. This same pathway was also reported in a transcriptome analysis from two cattle populations divergently selected for high and low RFI. We conclude that a GWAS pathway-based metaanalysis can be an appropriate method to uncover biological insights into RFI by combining useful information from different studies.

Rhipicephalus (Boophilus) microplus is the main cattle ectoparasite in tropical areas. Gir × Holstein crossbred cows are well adapted to different production systems in Brazil. In this context, we performed genome-wide association study (GWAS) and post-GWAS analyses for R. microplus resistance in an experimental Gir × Holstein F population. Single nucleotide polymorphisms (SNP) identified in GWAS were used to build gene networks and to investigate the breed of origin for its alleles. Tick artificial infestations were performed during the dry and rainy seasons. Illumina BovineSNP50 BeadChip (Illumina Inc., San Diego, CA) and single-step BLUP procedure was used for GWAS. Post-GWAS analyses were performed by gene ontology terms enrichment and gene transcription factors networks, generated from enriched transcription factors, identified from the promoter sequences of selected gene sets. The genetic origin of marker alleles in the F population was assigned using the breed of origin of alleles approach. Heritability estimates for tick counts were 0.40 ± 0.11 in the rainy season and 0.54 ± 0.11 in the dry season. The top ten 0.5-Mbp windows with the highest percentage of genetic variance explained by SNP markers were found in chromosomes 10 and 23 for both the dry and rainy seasons. Gene network analyses allowed the identification of genes involved with biological processes relevant to immune system functions (TREM1, TREM2, and CD83). Gene-transcription factors network allowed the identification of genes involved with immune functions (MYO5A, TREML1, and PRSS16). In resistant animals, the average proportion of animals showing significant SNPs with paternal and maternal alleles originated from Gir breed was 44.8% whereas the proportion of animals with both paternal and maternal alleles originated from Holstein breed was 11.3%. Susceptible animals showing both paternal and maternal alleles originated from Holstein breed represented 44.6% on average, whereas both paternal and maternal alleles originated from Gir breed animals represented 9.3%. This study allowed us to identify candidate genes for tick resistance in Gir × Holstein crossbreds in both rainy and dry seasons. According to the origin of alleles analysis, we found that most animals classified as resistant showed 2 alleles from Gir breed, while the susceptible ones showed alleles from Holstein. Based on these results, the identified genes may be thoroughly investigated in additional experiments aiming to validate their effects on tick resistance phenotype in cattle.

Meta-analysis based on a random-effects model is used to summarise and overcome the variability between divergent parameter estimates. We proposed a meta-analysis of published heritability and genetic-correlation estimates for reproduction, growth and carcass traits in purebred Nellore cattle. In total, 197 heritability and 107 genetic-correlation estimates from 62 scientific publications were used here. Most of traits (gestation length; weights at birth, 120, 210, 365 and 550 days of age; mature weight and all carcass traits) presented direct heritability estimates ranging from 0.20 to 0.40. Age at first calving presented the lowest value among direct heritabilities (0.1498); whereas the higher values (>0.40) were found for scrotal circumference at different ages and for weight at 450 days of age. Low maternal heritability estimates (ranging from 0.06 to 0.11) were observed for all growth traits. With the exception of correlation estimates involving the age at first calving, all other correlations were positive. High correlations (>0.85) were found mainly for the same trait at different ages. The results reported here will give support to genetic evaluations when reliable estimates for different traits in purebred Nellore cattle are not available.

The level of genetic diversity in a population is inversely proportional to the linkage disequilibrium (LD) between individual single nucleotide polymorphisms (SNPs) and quantitative trait loci (QTLs), leading to lower predictive ability of genomic breeding values (GEBVs) in high genetically diverse populations. Haplotype-based predictions could outperform individual SNP predictions by better capturing the LD between SNP and QTL. Therefore, we aimed to evaluate the accuracy and bias of individual-SNP- and haplotype-based genomic predictions under the single-step-genomic best linear unbiased prediction (ssGBLUP) approach in genetically diverse populations. We simulated purebred and composite sheep populations using literature parameters for moderate and low heritability traits. The haplotypes were created based on LD thresholds of 0.1, 0.3, and 0.6. Pseudo-SNPs from unique haplotype alleles were used to create the genomic relationship matrix (G) in the ssGBLUP analyses. Alternative scenarios were compared in which the pseudo-SNPs were combined with non-LD clustered SNPs, only pseudo-SNPs, or haplotypes fitted in a second G (two relationship matrices). The GEBV accuracies for the moderate heritability-trait scenarios fitting individual SNPs ranged from 0.41 to 0.55 and with haplotypes from 0.17 to 0.54 in the most (Ne ≅ 450) and less (Ne &lt; 200) genetically diverse populations, respectively, and the bias fitting individual SNPs or haplotypes ranged between −0.14 and −0.08 and from −0.62 to −0.08, respectively. For the low heritability-trait scenarios, the GEBV accuracies fitting individual SNPs ranged from 0.24 to 0.32, and for fitting haplotypes, it ranged from 0.11 to 0.32 in the more (Ne ≅ 250) and less (Ne ≅ 100) genetically diverse populations, respectively, and the bias ranged between −0.36 and −0.32 and from −0.78 to −0.33 fitting individual SNPs or haplotypes, respectively. The lowest accuracies and largest biases were observed fitting only pseudo-SNPs from blocks constructed with an LD threshold of 0.3 (p &lt; 0.05), whereas the best results were obtained using only SNPs or the combination of independent SNPs and pseudo-SNPs in one or two G matrices, in both heritability levels and all populations regardless of the level of genetic diversity. In summary, haplotype-based models did not improve the performance of genomic predictions in genetically diverse populations.