Development of genomic predictions for harvest and carcass weight in channel catfish

Characterization of Kiss/Kissr system and expression profiling through developmental stages indicate kiss1 to be the active isotype in Clarias magur

Kisspeptin (Kiss) and kisspeptin receptor (Kissr) system is a key regulator of GnRH expression in several vertebrates. The Indian catfish, Clarias magur, is popular in the Indian sub-continent, and a neo-type of the Asian catfish, C. batrachus. Catfish breeding is constrained as males do not release milt captivity with/without stimulation. Magur Kiss/Kissr system comprising of kiss1, kiss2, kissr1, and kissr2 genes was characterized for the first time. Full-length mRNA was sequenced using RACE PCR. Neighbor-joining tree of predicted proteins shows one clade of teleost orthologs. Magur whole genome (NCBI GenBank) has single copies of each gene, though yet unannotated/misannotated. Anomalies in the nomenclature of earlier sequences in GenBank were noted. Relative gene expression was profiled during various ontogenic stages, in six tissues including brain and gonads at maturity, and also in brains and gonads of premature and spent fish. Expression of gnrh1, gnrhr1, and gnrhr2 was estimated concomitantly. The kiss1 was the first to be twofold upregulated (P < 0.05) at 12 h post fertilization. Kiss/Kissr genes expressed primarily in the brain, ovary, and testis. Though kiss2 was 10 times higher than kiss1, only kiss1 showed significant modulation across stages and appears to be the active isotype that regulates GnRH in magur.

Weighted single-step genome-wide association study for direct and maternal genetic effects associated with birth and weaning weights in sheep

Body weight is an important economic trait for sheep meat production, and its genetic improvement is considered one of the main goals in the sheep breeding program. Identifying genomic regions that are associated with growth-related traits accelerates the process of animal breeding through marker-assisted selection, which leads to increased response to selection. In this study, we conducted a weighted single-step genome-wide association study (WssGWAS) to identify potential candidate genes for direct and maternal genetic effects associated with birth weight (BW) and weaning weight (WW) in Baluchi sheep. The data used in this research included 13,408 birth and 13,170 weaning records collected at Abbas-Abad Baluchi Sheep Breeding Station, Mashhad-Iran. Genotypic data of 94 lambs genotyped by Illumina 50K SNP BeadChip for 54,241 markers were used. The proportion of variance explained by genomic windows was calculated by summing the variance of SNPs within 1 megabase (Mb). The top 10 window genomic regions explaining the highest percentages of additive and maternal genetic variances were selected as candidate window genomic regions associated with body weights. Our findings showed that for BW, the top-ranked genomic regions (1 Mb windows) explained 4.30 and 4.92% of the direct additive and maternal genetic variances, respectively. The direct additive genetic variance explained by the genomic window regions varied from 0.31 on chromosome 1 to 0.59 on chromosome 8. The highest (0.84%) and lowest (0.32%) maternal genetic variances were explained by genomic windows on chromosome 10 and 17, respectively. For WW, the top 10 genomic regions explained 6.38 and 5.76% of the direct additive and maternal genetic variances, respectively. The highest and lowest contribution of direct additive genetic variances were 1.37% and 0.42%, respectively, both explained by genomic regions on chromosome 2. For maternal effects on WW, the highest (1.38%) and lowest (0.41%) genetic variances were explained by genomic windows on chromosome 2. Further investigation of these regions identified several possible candidate genes associated with body weight. Gene ontology analysis using the DAVID database identified several functional terms, such as translation repressor activity, nucleic acid binding, dehydroascorbic acid transporter activity, growth factor activity and SH2 domain binding.

Genetic diversity patterns in farmed rainbow trout (Oncorhynchus mykiss) populations using genome-wide SNP and haplotype data

Rainbow trout is one of the most popular aquaculture species worldwide, with a long history of domestication. However, limited information exists about the genetic diversity of farmed rainbow trout populations globally, with most available reports relying on low-throughput genotyping technologies. Notably, no information exists about the genetic diversity status of farmed rainbow trout in Sweden. Double-digest restriction-site-associated DNA sequencing was performed on more than 500 broodfish from two leading producers in Sweden and from the country's national breeding program. Following the detection of single nucleotide polymorphisms (SNPs), genetic diversity was studied by using either individual SNPs (n = 8680; one SNP retained per 300 bp sequence reads) or through SNP haplotypes (n = 20 558; all SNPs retained in 300 bp sequence reads). Similar amounts of genetic diversity were found amongst the three populations when individual SNPs were used. Furthermore, principal component analysis and discriminant analysis of principal components suggested two genetic clusters with the two industry populations grouped together. Genetic differentiation based on the FST fixation index was ~0.01 between the industry populations and ~0.05 when those were compared with the breeding program. Preliminary estimates of effective population size (Ne ) and inbreeding (based on runs of homozygosity; FROH ) were similar amongst the three populations (Ne  ≈ 50-80; median FROH  ≈ 0.11). Finally, the haplotype-based analysis suggested that animals from the breeding program had higher shared coancestry levels than those from the other two populations. Overall, our study provides novel insights into the genetic diversity and structure of Sweden's three main farmed rainbow trout populations, which could guide their future management.

Potential negative effects of genomic selection

Initial findings on genomic selection (GS) indicated substantial improvement for major traits, such as performance, and even successful selection for antagonistic traits. However, recent unofficial reports indicate an increased frequency of deterioration of secondary traits. This phenomenon may arise due to the mismatch between the accelerated selection process and resource allocation. Traits explicitly or implicitly accounted for by a selection index move toward the desired direction, whereas neglected traits change according to the genetic correlations with selected traits. Historically, the first stage of commercial genetic selection focused on production traits. After long-term selection, production traits improved, whereas fitness traits deteriorated, although this deterioration was partially compensated for by constantly improving management. Adding these fitness traits to the breeding objective and the used selection index also helped offset their decline while promoting long-term gains. Subsequently, the trend in observed fitness traits was a combination of a negative response due to genetic antagonism, positive response from inclusion in the selection index, and a positive effect of improving management. Under GS, the genetic trends accelerate, especially for well-recorded higher heritability traits, magnifying the negatively correlated responses for fitness traits. Then, the observed trend for fitness traits can become negative, especially because management modifications do not accelerate under GS. Additional deterioration can occur due to the rapid turnover of GS, as heritabilities for production traits can decline and the genetic antagonism between production and fitness traits can intensify. If the genetic parameters are not updated, the selection index will be inaccurate, and the intended gains will not occur. While the deterioration can accelerate for unrecorded or sparsely recorded fitness traits, GS can lead to an improvement for widely recorded fitness traits. In the context of GS, it is crucial to look for unexpected changes in relevant traits and take rapid steps to prevent further declines, especially in secondary traits. Changes can be anticipated by investigating the temporal dynamics of genetic parameters, especially genetic correlations. However, new methods are needed to estimate genetic parameters for the last generation with large amounts of genomic data.

First Genomic Prediction of Single-Step Models in Large Yellow Croaker

Genome selection is mainly used in disease-resistant traits of aquatic species; however, its implementation is hindered by a high cost of genotype and phenotype data collection. Single-step genomic best linear unbiased prediction (SSGBLUP) can integrate phenotypes, genetic markers, and pedigree records into simultaneous prediction without increasing genotyping costs. The objective of this study is to investigate the performance of SSGBLUP in large yellow croaker and to evaluate the effects of the number of phenotypic records and genotyping per family on the predictive ability of SSGBLUP. A large yellow croaker population consists of 6898 individuals from 14 families with survival time resistant against Cryptocaryon irritans (C. irritans), body weight (BW), and body length (BL) traits were collected, of which 669 individuals were genotyped. Results showed that the mean predictive ability of all traits in the individuals randomly sampling for SSGBLUP, GBLUP, and BLUP was 0.738, 0.738, and 0.736, respectively. Moreover, the predictive ability of SSGBLUP and BLUP models did not increase with the extra phenotypic records per family, in which the predictive ability of SSGBLUP and BLUP in survival time was 0.853 and 0.851 for only genotyped data (N = 0) used, and 0.852, 0.845 for all phenotypic records (N = 600) used, respectively. However, with the increase in the genotype number of training set, the prediction ability of SSGBLUP and GBLUP model was increased and the highest predictive ability was gained when the genotype number per family was 40 or 45. In addition, the prediction ability of SSGBLUP model was higher than that of GBLUP. Our study showed that the SSGBLUP model still has great potential and advantages in genomic breeding of large yellow croakers. It is recommended that each family provide 100 phenotypic individuals, of which 40 individuals with genotyping data for SSGBLUP model prediction and family resistance evaluation.

Understanding the phenotypic and genetic background of the lactose content in Sarda dairy sheep

Lactose, the principal carbohydrate found in milk, plays an important role in the physiological processes of milk production because it is related to milk volume, and it is responsible for the osmotic equilibrium between blood and milk in the mammary gland. In this study, factors affecting lactose content (LC) in sheep milk are investigated. For this purpose, 2,358 test-day records were sampled from 509 ewes (3-7 records per animal). The LC and other main milk traits were analyzed using a mixed linear model that included days in milk (DIM) class, parity, lambing month, and type of lambing as fixed effects and animal, permanent environment, and flock test day as random effects. The pedigree-based approach was used to estimate the heritability and repeatability of LC. Moreover, the genomic background of LC was investigated through a GWAS. The LC was affected by all tested factors (i.e., DIM class, parity, lambing month, and type of lambing). Low heritability (0.10 ± 0.05) and moderate repeatability (0.42 ± 0.02) were estimated for LC. High negative genetic correlations were estimated between LC and NaCl (-0.99 ± 0.01) and between LC and somatic cell count (-0.94 ± 0.05). Only 2 markers passed the chromosome-wide Bonferroni threshold. Results of the present study, although obtained on a relatively small sample, suggest the possibility to include LC in the breeding programs, particularly because of its strong relationship with NaCl and somatic cell count.

Genomic selection models substantially improve the accuracy of genetic merit predictions for fillet yield and body weight in rainbow trout using a multi-trait model and multi-generation progeny testing

BACKGROUND

In aquaculture, the proportion of edible meat (FY = fillet yield) is of major economic importance, and breeding animals of superior genetic merit for this trait can improve efficiency and profitability. Achieving genetic gains for fillet yield is possible using a pedigree-based best linear unbiased prediction (PBLUP) model with direct and indirect selection. To investigate the feasibility of using genomic selection (GS) to improve FY and body weight (BW) in rainbow trout, the prediction accuracy of GS models was compared to that of PBLUP. In addition, a genome-wide association study (GWAS) was conducted to identify quantitative trait loci (QTL) for the traits. All analyses were performed using a two-trait model with FY and BW, and variance components, heritability, and genetic correlations were estimated without genomic information. The data used included 14,165 fish in the pedigree, of which 2742 and 12,890 had FY and BW phenotypic records, respectively, and 2484 had genotypes from the 57K single nucleotide polymorphism (SNP) array.

RESULTS

The heritabilities were moderate, at 0.41 and 0.33 for FY and BW, respectively. Both traits were lowly but positively correlated (genetic correlation; r = 0.24), which suggests potential favourable correlated genetic gains. GS models increased prediction accuracy compared to PBLUP by up to 50% for FY and 44% for BW. Evaluations were found to be biased when validation was performed on future performances but not when it was performed on future genomic estimated breeding values.

CONCLUSIONS

The low but positive genetic correlation between fillet yield and body weight indicates that some improvement in fillet yield may be achieved through indirect selection for body weight. Genomic information increases the prediction accuracy of breeding values and is an important tool to accelerate genetic progress for fillet yield and growth in the current rainbow trout population. No significant QTL were found for either trait, indicating that both traits are polygenic, and that marker-assisted selection will not be helpful to improve these traits in this population.

Prediction ability for growth and maternal traits using SNP arrays based on different marker densities in Nellore cattle using the ssGBLUP

This study aimed to investigate the prediction ability for growth and maternal traits using different low-density customized SNP arrays selected by informativeness and distribution of markers across the genome employing single-step genomic BLUP (ssGBLUP). Phenotypic records for adjusted weight at 210 and 450 days of age were utilized. A total of 945 animals were genotyped with high-density chip, and 267 individuals born after 2008 were selected as validation population. We evaluated 11 scenarios using five customized density arrays (40 k, 20 k, 10 k, 5 k and 2 k) and the HD array was used as desirable scenario. The GEBV predictions and BIF (Beef Improvement Federation) accuracy were obtained with BLUPF90 family programs. Linear regression was used to evaluate the prediction ability, inflation, and bias of GEBV of each customized array. An overestimation of partial GEBVs in contrast with complete GEBVs and increase of BIF accuracy with the density arrays diminished were observed. For all traits, the prediction ability was higher as the array density increased and it was similar with customized arrays higher than 10 k SNPs. Level of inflation was lower as the density array increased of and was higher for MW210 effect. The bias was susceptible to overestimation of GEBVs when the density customized arrays decreased. These results revealed that the BIF accuracy is sensible to overestimation using low-density customized arrays while the prediction ability with least 10,000 informative SNPs obtained from the Illumina BovineHD BeadChip shows accurate and less biased predictions. Low-density customized arrays under ssGBLUP method could be feasible and cost-effective in genomic selection.

Genomic Predictions of Phenotypes and Pseudo-Phenotypes for Viral Nervous Necrosis Resistance, Cortisol Concentration, Antibody Titer and Body Weight in European Sea Bass

Simple Summary

Selective breeding programs based on genomic data are still not a common practice in aquaculture, although genomic selection has been widely demonstrated to be advantageous when trait phenotyping is a difficult task. In this study, we investigated the accuracy of predicting the phenotype and the estimated breeding value (EBV) of three Bayesian models and a Random Forest algorithm exploiting the information of a genome-wide SNP panel for European sea bass. The genomic predictions were developed for mortality caused by viral nervous necrosis, post-stress cortisol concentration, antibody titer against nervous necrosis virus and body weight. Selective breeding based on genomic data is a possible option for improving these traits while overcoming difficulties related to individual phenotyping of the investigated traits. Our results evidenced that the EBV used as a pseudo-phenotype enhances the predictive performances of genomic models, and that EBV can be predicted with satisfactory accuracy. The genomic prediction of the EBV for mortality might also be used to classify the phenotype for the same trait.

Abstract

In European sea bass (Dicentrarchus labrax L.), the viral nervous necrosis mortality (MORT), post-stress cortisol concentration (HC), antibody titer (AT) against nervous necrosis virus and body weight (BW) show significant heritability, which makes selective breeding a possible option for their improvement. An experimental population (N = 650) generated by a commercial broodstock was phenotyped for the aforementioned traits and genotyped with a genome-wide SNP panel (16,075 markers). We compared the predictive accuracies of three Bayesian models (Bayes B, Bayes C and Bayesian Ridge Regression) and a machine-learning method (Random Forest). The prediction accuracy of the EBV for MORT was approximately 0.90, whereas the prediction accuracies of the EBV and the phenotype were 0.86 and 0.21 for HC, 0.79 and 0.26 for AT and 0.71 and 0.38 for BW. The genomic prediction of the EBV for MORT used to classify the phenotype for the same trait showed moderate classification performance. Genome-wide association studies confirmed the polygenic nature of MORT and demonstrated a complex genetic structure for HC and AT. Genomic predictions of the EBV for MORT could potentially be used to classify the phenotype of the same trait, though further investigations on a larger experimental population are needed.

Analysis of Autozygosity Using Whole-Genome Sequence Data of Full-Sib Families in Pikeperch (Sander lucioperca)

Pikeperch (Sander lucioperca) has emerged as a high value species to the aquaculture industry. However, its farming techniques are at an early stage and its production is often performed without a selective breeding program, potentially leading to high levels of inbreeding. In this study, we identified and characterized autozygosity based on genome-wide runs of homozygosity (ROH) on a sample of parental and offspring individuals, determined effective population size (N_e), and assessed relatedness among parental individuals. A mean of 2,235 ± 526 and 1,841 ± 363 ROH segments per individual, resulting in a mean inbreeding coefficient of 0.33 ± 0.06 and 0.25 ± 0.06 were estimated for the progeny and parents, respectively. N_e was about 12 until four generations ago and at most 106 for 63 generations in the past, with varying genetic relatedness amongst the parents. This study shows the importance of genomic information when family relationships are unknown and the need of selective breeding programs for reproductive management decisions in the aquaculture industry.

Genomic Selection in Aquaculture Species

To date, genomic prediction has been conducted in about 20 aquaculture species, with a preference for intra-family genomic selection (GS). For every trait under GS, the increase in accuracy obtained by genomic estimated breeding values instead of classical pedigree-based estimation of breeding values is very important in aquaculture species ranging from 15% to 89% for growth traits, and from 0% to 567% for disease resistance. Although the implementation of GS in aquaculture is of little additional investment in breeding programs already implementing sib testing on pedigree, the deployment of GS remains sparse, but could be boosted by adaptation of cost-effective imputation from low-density panels. Moreover, GS could help to anticipate the effect of climate change by improving sustainability-related traits such as production yield (e.g., carcass or fillet yields), feed efficiency or disease resistance, and by improving resistance to environmental variation (tolerance to temperature or salinity variation). This chapter synthesized the literature in applications of GS in finfish, crustaceans and molluscs aquaculture in the present and future breeding programs.

Genome-wide association study on milk production and somatic cell score for Thai dairy cattle using weighted single-step approach with random regression test-day model

Genome-wide association studies are a powerful tool to identify genomic regions and variants associated with phenotypes. However, only limited mutual confirmation from different studies is available. The objectives of this study were to identify genomic regions as well as genes and pathways associated with the first-lactation milk, fat, protein, and total solid yields; fat, protein, and total solid percentage; and somatic cell score (SCS) in a Thai dairy cattle population. Effects of SNPs were estimated by a weighted single-step GWAS, which back-solved the genomic breeding values predicted using single-step genomic BLUP (ssGBLUP) fitting a single-trait random regression test-day model. Genomic regions that explained at least 0.5% of the total genetic variance were selected for further analyses of candidate genes. Despite the small number of genotyped animals, genomic predictions led to an improvement in the accuracy over the traditional BLUP. Genomic predictions using weighted ssGBLUP were slightly better than the ssGBLUP. The genomic regions associated with milk production traits contained 210 candidate genes on 19 chromosomes [Bos taurus autosome (BTA) 1 to 7, 9, 11 to 16, 20 to 21, 26 to 27 and 29], whereas 21 candidate genes on 3 chromosomes (BTA 11, 16, and 21) were associated with SCS. Many genomic regions explained a small fraction of the genetic variance, indicating polygenic inheritance of the studied traits. Several candidate genes coincided with previous reports for milk production traits in Holstein cattle, especially a large region of genes on BTA14. We identified 141 and 5 novel genes related to milk production and SCS, respectively. These novel genes were also found to be functionally related to heat tolerance (e.g., SLC45A2, IRAG1, and LOC101902172), longevity (e.g., SYT10 and LOC101903327), and fertility (e.g., PAG1). These findings may be attributed to indirect selection in our population. Identified biological networks including intracellular cell transportation and protein catabolism implicate milk production, whereas the immunological pathways such as lymphocyte activation are closely related to SCS. Further studies are required to validate our findings before exploiting them in genomic selection.

Genomic prediction for testes weight of the tiger pufferfish, Takifugu rubripes, using medium to low density SNPs

Aquaculture production is expected to increase with the help of genomic selection (GS). The possibility of performing GS using only a small number of SNPs has been examined in order to reduce genotyping costs; however, the practicality of this approach is still unclear. Here, we tested whether the effects of reducing the number of SNPs impaired the prediction accuracy of GS for standard length, body weight, and testes weight in the tiger pufferfish (Takifugu rubripes). High values for predictive ability (0.563-0.606) were obtained with 4000 SNPs for all traits under a genomic best linear unbiased predictor (GBLUP) model. These values were still within an acceptable range with 1200 SNPs (0.554-0.588). However, predictive abilities and prediction accuracies deteriorated using less than 1200 SNPs largely due to the reduced power in accurately estimating the genetic relationship among individuals; family structure could still be resolved with as few as 400 SNPs. This suggests that the SNPs informative for estimation of genetic relatedness among individuals differ from those for inference of family structure, and that non-random SNP selection based on the effects on family structure (e.g., site-F_ST, principal components, or random forest) is unlikely to increase the prediction accuracy for these traits.

Haplotype genomic prediction of phenotypic values based on chromosome distance and gene boundaries using low-coverage sequencing in Duroc pigs

Background

Genomic selection using single nucleotide polymorphism (SNP) markers has been widely used for genetic improvement of livestock, but most current methods of genomic selection are based on SNP models. In this study, we investigated the prediction accuracies of haplotype models based on fixed chromosome distances and gene boundaries compared to those of SNP models for genomic prediction of phenotypic values. We also examined the reasons for the successes and failures of haplotype genomic prediction.

Methods

We analyzed a swine population of 3195 Duroc boars with records on eight traits: body judging score (BJS), teat number (TN), age (AGW), loin muscle area (LMA), loin muscle depth (LMD) and back fat thickness (BF) at 100 kg live weight, and average daily gain (ADG) and feed conversion rate (FCR) from 30 to100 kg live weight. Ten-fold validation was used to evaluate the prediction accuracy of each SNP model and each multi-allelic haplotype model based on 488,124 autosomal SNPs from low-coverage sequencing. Haplotype blocks were defined using fixed chromosome distances or gene boundaries.

Results

Compared to the best SNP model, the accuracy of predicting phenotypic values using a haplotype model was greater by 7.4% for BJS, 7.1% for AGW, 6.6% for ADG, 4.9% for FCR, 2.7% for LMA, 1.9% for LMD, 1.4% for BF, and 0.3% for TN. The use of gene-based haplotype blocks resulted in the best prediction accuracy for LMA, LMD, and TN. Compared to estimates of SNP additive heritability, estimates of haplotype epistasis heritability were strongly correlated with the increase in prediction accuracy by haplotype models. The increase in prediction accuracy was largest for BJS, AGW, ADG, and FCR, which also had the largest estimates of haplotype epistasis heritability, 24.4% for BJS, 14.3% for AGW, 14.5% for ADG, and 17.7% for FCR. SNP and haplotype heritability profiles across the genome identified several genes with large genetic contributions to phenotypes: NUDT3 for LMA, LMD and BF, VRTN for TN, COL5A2 for BJS, BSND for ADG, and CARTPT for FCR.

Conclusions

Haplotype prediction models improved the accuracy for genomic prediction of phenotypes in Duroc pigs. For some traits, the best prediction accuracy was obtained with haplotypes defined using gene regions, which provides evidence that functional genomic information can improve the accuracy of haplotype genomic prediction for certain traits.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12711-021-00661-y.

Genomic prediction of milk-production traits and somatic cell score using single-step genomic best linear unbiased predictor with random regression test-day model in Thai dairy cattle

Cow genotypes are expected to improve the accuracy of genomic estimated breeding values (GEBV) for young bulls in relatively small populations such as Thai Holstein-Friesian crossbred dairy cattle in Thailand. The objective of this study was to investigate the effect of cow genotypes on the predictive ability and individual accuracies of GEBV for young dairy bulls in Thailand. Test-day data included milk yield (n = 170,666), milk component traits (fat yield, protein yield, total solids yield, fat percentage, protein percentage, and total solids percentage; n = 160,526), and somatic cell score (n = 82,378) from 23,201, 82,378, and 13,737 (for milk yield, milk component traits, and SCS, respectively) cows calving between 1993 and 2017, respectively. Pedigree information included 51,128; 48,834; and 32,743 animals for milk yield, milk component traits, and somatic cell score, respectively. Additionally, 876, 868, and 632 pedigreed animals (for milk yield, milk component traits, and SCS, respectively) were genotyped (152 bulls and 724 cows), respectively, using Illumina Bovine SNP50 BeadChip. We cut off the data in the last 6 yr, and the validation animals were defined as genotyped bulls with no daughters in the truncated set. We calculated GEBV using a single-step random regression test-day model (SS-RR-TDM), in comparison with estimated breed value (EBV) based on the pedigree-based model used as the official method in Thailand (RR-TDM). Individual accuracies of GEBV were obtained by inverting the coefficient matrix of the mixed model equations, whereas validation accuracies were measured by the Pearson correlation between deregressed EBV from the full data set and (G)EBV predicted with the reduced data set. When only bull genotypes were used, on average, SS-RR-TDM increased individual accuracies by 0.22 and validation accuracies by 0.07, compared with RR-TDM. With cow genotypes, the additional increase was 0.02 for individual accuracies and 0.06 for validation accuracies. The inflation of GEBV tended to be reduced using cow genotypes. Genomic evaluation by SS-RR-TDM is feasible to select young bulls for the longitudinal traits in Thai dairy cattle, and the accuracy of selection is expected to be increased with more genotypes. Genomic selection using the SS-RR-TDM should be implemented in the routine genetic evaluation of the Thai dairy cattle population. The genetic evaluation should consider including genotypes of both sires and cows.

Genomic Selection and Genome-wide Association Study for Feed-Efficiency Traits in a Farmed Nile Tilapia (Oreochromis niloticus) Population

Nile tilapia is a key aquaculture species with one of the highest production volumes globally. Genetic improvement of feed efficiency via selective breeding is an important goal, and genomic selection may expedite this process. The aims of this study were to 1) dissect the genetic architecture of feed-efficiency traits in a Nile tilapia breeding population, 2) map the genomic regions associated with these traits and identify candidate genes, 3) evaluate the accuracy of breeding value prediction using genomic data, and 4) assess the impact of the genetic marker density on genomic prediction accuracies. Using an experimental video recording trial, feed conversion ratio (FCR), body weight gain (BWG), residual feed intake (RFI) and feed intake (FI) traits were recorded in 40 full-sibling families from the GIFT (Genetically Improved Farmed Tilapia) Nile tilapia breeding population. Fish were genotyped with a ThermoFisher Axiom 65 K Nile tilapia SNP array. Significant heritabilities, ranging from 0.12 to 0.22, were estimated for all the assessed traits using the genomic relationship matrix. A negative but favourable genetic correlation was found between BWG and the feed-efficiency related traits; -0.60 and -0.63 for FCR and RFI, respectively. While the genome-wide association analyses suggested a polygenic genetic architecture for all the measured traits, there were significant QTL identified for BWG and FI on chromosomes seven and five respectively. Candidate genes previously found to be associated with feed-efficiency traits were located in these QTL regions, including ntrk3a, ghrh and eif4e3. The accuracy of breeding value prediction using the genomic data was up to 34% higher than using pedigree records. A SNP density of approximately 5,000 SNPs was sufficient to achieve similar prediction accuracy as the full genotype data set. Our results highlight the potential of genomic selection to improve feed efficiency traits in Nile tilapia breeding programmes.

Impact of genotypic errors with equal and unequal family contribution on accuracy of genomic prediction in aquaculture using simulation

Genotypic errors, conflict between recorded genotype and the true genotype, can lead to false or biased population genetic parameters. Here, the effect of genotypic errors on accuracy of genomic predictions and genomic relationship matrix are investigated using a simulation study based on population and genomic structure comparable to black tiger prawn, Penaeus monodon. Fifty full-sib families across five generations with phenotypic and genotypic information on 53 K SNPs were simulated. Ten replicates of different scenarios with three heritability estimates, equal and unequal family contributions were generated. Within each scenario, four SNP densities and three genotypic error rates in each SNP density were implemented. Results showed that family contribution did not have a substantial impact on accuracy of predictions across different datasets. In the absence of genotypic errors, 3 K SNP density was found to be efficient in estimating the accuracy, whilst increasing the SNP density from 3 to 20 K resulted in a marginal increase in accuracy of genomic predictions using the current population and genomic parameters. In addition, results showed that the presence of even 10% errors in a 10 and 20 K SNP panel might not have a severe impact on accuracy of predictions. However, below 10 K marker density, even a 5% error can result in lower accuracy of predictions.

Metafounders May Reduce Bias in Composite Cattle Genomic Predictions

Metafounders are pseudo-individuals that act as proxies for animals in base populations. When metafounders are used, individuals from different breeds can be related through pedigree, improving the compatibility between genomic and pedigree relationships. The aim of this study was to investigate the use of metafounders and unknown parent groups (UPGs) for the genomic evaluation of a composite beef cattle population. Phenotypes were available for scrotal circumference at 14 months of age (SC14), post weaning gain (PWG), weaning weight (WW), and birth weight (BW). The pedigree included 680,551 animals, of which 1,899 were genotyped for or imputed to around 30,000 single-nucleotide polymorphisms (SNPs). Evaluations were performed based on pedigree (BLUP), pedigree with UPGs (BLUP_UPG), pedigree with metafounders (BLUP_MF), single-step genomic BLUP (ssGBLUP), ssGBLUP with UPGs for genomic and pedigree relationship matrices (ssGBLUP_UPG) or only for the pedigree relationship matrix (ssGBLUP_UPGA), and ssGBLUP with metafounders (ssGBLUP_MF). Each evaluation considered either four or 10 groups that were assigned based on breed of founders and intermediate crosses. To evaluate model performance, we used a validation method based on linear regression statistics to obtain accuracy, stability, dispersion, and bias of (genomic) estimated breeding value [(G)EBV]. Overall, relationships within and among metafounders were stronger in the scenario with 10 metafounders. Accuracy was greater for models with genomic information than for BLUP. Also, the stability of (G)EBVs was greater when genomic information was taken into account. Overall, pedigree-based methods showed lower inflation/deflation (regression coefficients close to 1.0) for SC14, WWM, and BWD traits. The level of inflation/deflation for genomic models was small and trait-dependent. Compared with regular ssGBLUP, ssGBLUP_MF4 displayed regression coefficient closer to one SC14, PWG, WWM, and BWD. Genomic models with metafounders seemed to be slightly more stable than models with UPGs based on higher similarity of results with different numbers of groups. Further, metafounders can help to reduce bias in genomic evaluations of composite beef cattle populations without reducing the stability of GEBVs.

Investigating the persistence of accuracy of genomic predictions over time in broilers

Accuracy of genomic predictions is an important component of the selection response. The objectives of this research were: 1) to investigate trends for prediction accuracies over time in a broiler population of accumulated phenotypes, genotypes, and pedigrees and 2) to test if data from distant generations are useful to maintain prediction accuracies in selection candidates. The data contained 820K phenotypes for a growth trait (GT), 200K for two feed efficiency traits (FE1 and FE2), and 42K for a carcass yield trait (CY). The pedigree included 1,252,619 birds hatched over 7 years, of which 154,318 from the last 4 years were genotyped. Training populations were constructed adding 1 year of data sequentially, persistency of accuracy over time was evaluated using predictions from birds hatched in the three generations following or in the years after the training populations. In the first generation, before genotypes became available for the training populations (first 3 years of data), accuracies remained almost stable with successive additions of phenotypes and pedigree to the accumulated dataset. The inclusion of 1 year of genotypes in addition to 4 years of phenotypes and pedigree in the training population led to increases in accuracy of 54% for GT, 76% for FE1, 110% for CY, and 38% for FE2; on average, 74% of the increase was due to genomics. Prediction accuracies declined faster without than with genomic information in the training populations. When genotypes were unavailable, the average decline in prediction accuracy across traits was 41% from the first to the second generation of validation, and 51% from the second to the third generation of validation. When genotypes were available, the average decline across traits was 14% from the first to the second generation of validation, and 3% from the second to the third generation of validation. Prediction accuracies in the last three generations were the same when the training population included 5 or 2 years of data, and a decrease of ~7% was observed when the training population included only 1 year of data. Training sets including genomic information provided an increase in accuracy and persistence of genomic predictions compared with training sets without genomic data. The two most recent years of pedigree, phenotypic, and genomic data were sufficient to maintain prediction accuracies in selection candidates. Similar conclusions were obtained using validation populations per year.

Investigating pig survival in different production phases using genomic models

Pig survival is an economically important trait with relevant social welfare implications, thus standing out as an important selection criterion for the current pig farming system. We aimed to estimate (co)variance components for survival in different production phases in a crossbred pig population as well as to investigate the benefit of including genomic information through single-step genomic best linear unbiased prediction (ssGBLUP) on the prediction accuracy of survival traits compared with results from traditional BLUP. Individual survival records on, at most, 64,894 crossbred piglets were evaluated under two multi-trait threshold models. The first model included farrowing, lactation, and combined postweaning survival, whereas the second model included nursery and finishing survival. Direct and maternal breeding values were estimated using BLUP and ssGBLUP methods. Furthermore, prediction accuracy, bias, and dispersion were accessed using the linear regression validation method. Direct heritability estimates for survival in all studied phases were low (from 0.02 to 0.08). Survival in preweaning phases (farrowing and lactation) was controlled by the dam and piglet additive genetic effects, although the maternal side was more important. Postweaning phases (nursery, finishing, and the combination of both) showed the same or higher direct heritabilities compared with preweaning phases. The genetic correlations between survival traits within preweaning and postweaning phases were favorable and strong, but correlations between preweaning and postweaning phases were moderate. The prediction accuracy of survival traits was low, although it increased by including genomic information through ssGBLUP compared with the prediction accuracy from BLUP. Direct and maternal breeding values were similarly accurate with BLUP, but direct breeding values benefited more from genomic information. Overall, a slight increase in bias was observed when genomic information was included, whereas dispersion of breeding values was greatly reduced. Combined postweaning survival presented higher direct heritability than in the preweaning phases and the highest prediction accuracy among all evaluated production phases, therefore standing out as a candidate trait for improving survival. Survival is a complex trait with low heritability; however, important genetic gains can still be obtained, especially under a genomic prediction framework.

A genomic‐based vision on the genetic diversity and key performance traits in selectively bred Arctic charr ( Salvelinus alpinus )

Routine implementation of genomic information for guiding selection decisions is not yet common in the majority of aquaculture species. Reduced representation sequencing approaches offer a cost‐effective solution for obtaining genome‐wide information in species with a limited availability of genomic resources. In the current study, we implemented double‐digest restriction site‐associated DNA sequencing (ddRAD‐seq) on an Arctic charr strain with the longest known history of selection (approximately 40 years) aiming to improve selection decisions. In total, 1730 animals reared at four different farms in Sweden and spanning from year classes 2013–2017 were genotyped using ddRAD‐seq. Approximately 5000 single nucleotide polymorphisms (SNPs) were identified, genetic diversity‐related metrics were estimated, and genome‐wide association studies (GWAS) for body length at different time points and age of sexual maturation were conducted. Low genetic differentiation amongst animals from the different farms was observed based on both the results from pairwise F_st values and principal component analysis (PCA). The existence of associations was investigated between the mean genome‐wide heterozygosity of each full‐sib family (year class 2017) and the corresponding inbreeding coefficient or survival to the eyed stage. A moderate correlation (−0.33) was estimated between the mean observed heterozygosity of each full‐sib family and the corresponding inbreeding coefficient, while no linear association was obtained with the survival to the eyed stage. GWAS did not detect loci with major effect for any of the studied traits. However, genomic regions explaining more than 1% of the additive genetic variance for either studied traits were suggested across 14 different chromosomes. Overall, key insights valuable for future selection decisions of Arctic charr have been obtained, suggesting ddRAD as an attractive genotyping platform for obtaining genome‐wide information in a cost‐effective manner.

Single-Step Genome Wide Association Study Identifies QTL Signals for Untrimmed and Trimmed Thigh Weight in Italian Crossbred Pigs for Dry-Cured Ham Production

Simple Summary

Along with the traditional traits, swine breeding programs for Italian dry-cured ham production have recently aimed to include novel phenotypes. The identification of the genomic regions underlying such new traits helps to untangle their genetic architecture and may provide useful information to be integrated in genetic selection. With this aim, we estimated genetic parameters and conducted a single step genome wide association studies (GWAS) on untrimmed and trimmed thigh weight considering two pig crossbred lines approved for Italian Protected Designation of Origin ham production. Quantitative trait loci (QTLs) were characterized based on the variance of 10-SNP sliding windows genomic estimated breeding values. In particular, we identified interesting QTL signals on several chromosomes, notably on chromosome 4, 6, 7 and 15. A high heritability and genetic correlation were observed for the two traits under investigation and although independent studies including other pig populations are required to disentangle the possible effects of specific linkage disequilibrium in our population, our findings suggest that such QTL could be investigated in future pig breeding programs to improve the reliability of genomic estimated breeding values for the dry-cured ham production.

Abstract

Protected Designation of Origin (PDO) dry-cured ham is the most important product in the Italian pig breeding industry, mainly oriented to produce heavy pig carcasses to obtain hams of the right weight and maturity. Recently, along with the traditional traits swine breeding programs have aimed to include novel carcass traits. The identification at the genome level of quantitative trait loci (QTLs) affecting such new traits helps to reveal their genetic determinism and may provide information to be integrated in prediction models in order to improve prediction accuracy as well as to identify candidate genes underlying such traits. This study aimed to estimate genetic parameters and perform a single step genome wide association studies (ssGWAS) on novel carcass traits such as untrimmed (UTW) and trimmed thigh weight (TTW) in two pig crossbred lines approved for the ham production of the Italian PDO. With this purpose, phenotypes were collected from ~1800 animals and 240 pigs were genotyped with Illumina PorcineSNP60 Beadchip. The single-step genomic BLUP procedure was used for the heritability estimation and to implement the ssGWAS. QTL were characterized based on the variance of 10-SNP sliding window genomic estimated breeding values. Moderate heritabilities were detected and QTL signals were identified on chromosome 1, 4, 6, 7, 11 and 15 for both traits. As expected, the genetic correlation among the two traits was very high (~0.99). The QTL regions encompassed a total of 249 unique candidate genes, some of which were already reported in association with growth, carcass or ham weight traits in pigs. Although independent studies are required to further verify our findings and disentangle the possible effects of specific linkage disequilibrium in our population, our results support the potential use of such new QTL information in future breeding programs to improve the reliability of genomic prediction.

Accounting for Population Structure and Phenotypes From Relatives in Association Mapping for Farm Animals: A Simulation Study

Population structure or genetic relatedness should be considered in genome association studies to avoid spurious association. The most used methods for genome-wide association studies (GWAS) account for population structure but are limited to genotyped individuals with phenotypes. Single-step GWAS (ssGWAS) can use phenotypes from non-genotyped relatives; however, its ability to account for population structure has not been explored. Here we investigate the equivalence among ssGWAS, efficient mixed-model association expedited (EMMAX), and genomic best linear unbiased prediction GWAS (GBLUP-GWAS), and how they differ from the single-SNP analysis without correction for population structure (SSA-NoCor). We used simulated, structured populations that mimicked fish, beef cattle, and dairy cattle populations with 1040, 5525, and 1,400 genotyped individuals, respectively. Larger populations were also simulated that had up to 10-fold more genotyped animals. The genomes were composed by 29 chromosomes, each harboring one QTN, and the number of simulated SNPs was 35,000 for the fish and 65,000 for the beef and dairy cattle populations. Males and females were genotyped in the fish and beef cattle populations, whereas only males had genotypes in the dairy population. Phenotypes for a trait with heritability varying from 0.25 to 0.35 were available in both sexes for the fish population, but only for females in the beef and dairy cattle populations. In the latter, phenotypes of daughters were projected into genotyped sires (i.e., deregressed proofs) before applying EMMAX and SSA-NoCor. Although SSA-NoCor had the largest number of true positive SNPs among the four methods, the number of false negatives was two–fivefold that of true positives. GBLUP-GWAS and EMMAX had a similar number of true positives, which was slightly smaller than in ssGWAS, although the difference was not significant. Additionally, no significant differences were observed when deregressed proofs were used as pseudo-phenotypes in EMMAX compared to daughter phenotypes in ssGWAS for the dairy cattle population. Single-step GWAS accounts for population structure and is a straightforward method for association analysis when only a fraction of the population is genotyped and/or when phenotypes are available on non-genotyped relatives.

Toward Genome-Based Selection in Asian Seabass: What Can We Learn From Other Food Fishes and Farm Animals?

Due to the steadily increasing need for seafood and the plateauing output of fisheries, more fish need to be produced by aquaculture production. In parallel with the improvement of farming methods, elite food fish lines with superior traits for production must be generated by selection programs that utilize cutting-edge tools of genomics. The purpose of this review is to provide a historical overview and status report of a selection program performed on a catadromous predator, the Asian seabass (Lates calcarifer, Bloch 1790) that can change its sex during its lifetime. We describe the practices of wet lab, farm and lab in detail by focusing onto the foundations and achievements of the program. In addition to the approaches used for selection, our review also provides an inventory of genetic/genomic platforms and technologies developed to (i) provide current and future support for the selection process; and (ii) improve our understanding of the biology of the species. Approaches used for the improvement of terrestrial farm animals are used as examples and references, as those processes are far ahead of the ones used in aquaculture and thus they might help those working on fish to select the best possible options and avoid potential pitfalls.

First genomic prediction and genome‐wide association for complex growth‐related traits in Rock Bream (Oplegnathus fasciatus)

Rock Bream (Oplegnathus fasciatus) is an important aquaculture species for offshore cage aquaculture and fish stocking of marine ranching in East Asia. Genomic selection has the potential to expedite genetic gain for the key target traits of a breeding program, but has not yet been evaluated in Oplegnathus. The purposes of the present study were to explore the performance of genomic selection to improve breeding value accuracy through real data analyses using six statistical models and to carry out genome‐wide association studies (GWAS) to dissect the genetic architecture of economically vital growth‐related traits (body weight, total length, and body depth) in the O. fasciatus population. After quality control, genotypes for 16,162 SNPs were acquired for 455 fish. Heritability was estimated to be moderate for the three traits (0.38 for BW, 0.33 for TL, and 0.24 for BD), and results of GWAS indicated that the underlying genetic architecture was polygenic. Six statistic models (GBLUP, BayesA, BayesB, BayesC, Bayesian Ridge‐Regression, and Bayesian LASSO) showed similar performance for the predictability of genomic estimated breeding value (GEBV). The low SNP density (around 1 K selected SNP based on GWAS) is sufficient for accurate prediction on the breeding value for the three growth‐related traits in the current studied population, which will provide a good compromise between genotyping costs and predictability in such standard breeding populations advanced. These consequences illustrate that the employment of genomic selection in O. fasciatus breeding could provide advantages for the selection of breeding candidates to facilitate complex economic growth traits.

Genomic information allows for more accurate breeding values for milkability in dual-purpose Italian Simmental cattle

Milkability is a trait related to the milking efficiency of an animal, and it is a component of the herd profitability. Due to its economic importance, milkability is currently included in the selection index of the Italian Simmental cattle breed with a weight of 7.5%. This lowly heritable trait is measured on a subjective scale from 1 to 3 (1 = slow, 3 = fast), and genetic evaluations are performed by pedigree-based BLUP. Genomic information is now available for some animals in the Italian Simmental population, and its inclusion in the genetic evaluation system could increase accuracy of breeding values and genetic progress for milkability. The aim of this study was to test the feasibility and advantages of having a genomic evaluation for this trait in the Italian Simmental population. Phenotypes were available for 131,308 cows. A total of 9,526 animals had genotypes for 42,152 loci; among the genotyped animals, 2,455 were cows with phenotypes, and the other were their relatives. The youngest cows with both phenotypes and genotypes (n = 900) were identified as selection candidates. Variance components and heritability were estimated using pedigree information, whereas genetic and genomic evaluations were carried out using BLUP and single-step genomic BLUP (ssGBLUP), respectively. In addition, a weighted ssGBLUP was assessed using genomic regions from a genome-wide association study. Evaluation models were validated using theoretical and realized accuracies. The estimated heritability for milkability was 0.12 ± 0.01. The mean theoretical accuracies for selection candidates were 0.43 ± 0.08 (BLUP) and 0.53 ± 0.06 (ssGBLUP). The mean realized accuracies based on linear regression statistics were 0.29 (BLUP) and 0.40 (ssGBLUP). No genomic regions were significantly associated with milkability, thus no improvements in accuracy were observed when using weighted ssGBLUP. Results indicated that genomic information could improve the accuracy of breeding values and increase genetic progress for milkability in Italian Simmental.

Genomic predictions for fillet yield and firmness in rainbow trout using reduced-density SNP panels

Background

One of the most important goals for the rainbow trout aquaculture industry is to improve fillet yield and fillet quality. Previously, we showed that a 50 K transcribed-SNP chip can be used to detect quantitative trait loci (QTL) associated with fillet yield and fillet firmness. In this study, data from 1568 fish genotyped for the 50 K transcribed-SNP chip and ~ 774 fish phenotyped for fillet yield and fillet firmness were used in a single-step genomic BLUP (ssGBLUP) model to compute the genomic estimated breeding values (GEBV). In addition, pedigree-based best linear unbiased prediction (PBLUP) was used to calculate traditional, family-based estimated breeding values (EBV).

Results

The genomic predictions outperformed the traditional EBV by 35% for fillet yield and 42% for fillet firmness. The predictive ability for fillet yield and fillet firmness was 0.19–0.20 with PBLUP, and 0.27 with ssGBLUP. Additionally, reducing SNP panel densities indicated that using 500–800 SNPs in genomic predictions still provides predictive abilities higher than PBLUP.

Conclusion

These results suggest that genomic evaluation is a feasible strategy to identify and select fish with superior genetic merit within rainbow trout families, even with low-density SNP panels.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12864-021-07404-9.

Assessing the potential of improving growth and survival to the eyed stage in selectively bred Arctic charr (Salvelinus alpinus)

The Arctic charr breeding programme has been a main driving force for developing the aquaculture industry in Sweden. Selection has been performed for almost 40 years using animals from a closed breeding nucleus. The aim of the current study was to evaluate the potential of further improving growth-related traits taking into account the existence of genotype-by-environment interaction. Furthermore, we investigated the magnitude of the genetic component associated with survival to the eyed stage and potential associations with inbreeding coefficients. A preliminary heritability estimate of 0.23 (SE 0.20) was obtained for survival to the eyed stage using records spanning from 2000 to 2017 (n = 230). Moreover, moderate-to-high heritability estimates (0.27-0.49) were obtained for growth-related traits (body weight and length), using animals from the latest generation of selection (year class 2017). Those animals (n = 2,776), originating from 55 full-sib families, were split into two groups and reared in separate land-based facilities of commercial fish farms in Sweden. The growth-related traits were recorded twice in both sites when animals were of >1 and >2 years of age. Existence of sexual growth dimorphism was indicated with the males having on average 6%-8% higher total length and 22%-34% higher body weight. Furthermore, high genetic correlations regarding growth traits were obtained amongst animals reared at the two different sites (0.82-0.95). In addition, we assessed the accuracy of best linear unbiased prediction (BLUP)-derived estimated breeding values (EBVs) when phenotypes from each rearing site were subsequently masked and used as a validation set. A mean prediction accuracy of 0.60 (length) and 0.64 (weight) were derived for both rearing sites. Overall, our results suggest that further growth improvements should be possible in the subsequent generations of selection. Finally, even though indications for the existence of an underlying genetic component(s) involved in survival to the eyed stage were obtained additional data will be required for elucidating its magnitude.

Validation of single-step GBLUP genomic predictions from threshold models using the linear regression method: An application in chicken mortality

The objective of this study was to determine whether the linear regression (LR) method could be used to validate genomic threshold models. Statistics for the LR method were computed from estimated breeding values (EBVs) using the whole and truncated data sets with variances from the reference and validation populations. The method was tested using simulated and real chicken data sets. The simulated data set included 10 generations of 4,500 birds each; genotypes were available for the last three generations. Each animal was assigned a continuous trait, which was converted to a binary score assuming an incidence of failure of 7%. The real data set included the survival status of 186,596 broilers (mortality rate equal to 7.2%) and genotypes of 18,047 birds. Both data sets were analysed using best linear unbiased predictor (BLUP) or single-step GBLUP (ssGBLUP). The whole data set included all phenotypes available, whereas in the partial data set, phenotypes of the most recent generation were removed. In the simulated data set, the accuracies based on the LR formulas were 0.45 for BLUP and 0.76 for ssGBLUP, whereas the correlations between true breeding values and EBVs (i.e. true accuracies) were 0.37 and 0.65, respectively. The gain in accuracy by adding genomic information was overestimated by 0.09 when using the LR method compared to the true increase in accuracy. However, when the estimated ratio between the additive variance computed based on pedigree only and on pedigree and genomic information was considered, the difference between true and estimated gain was <0.02. Accuracies of BLUP and ssGBLUP with the real data set were 0.41 and 0.47, respectively. This small improvement in accuracy when using ssGBLUP with the real data set was due to population structure and lower heritability. The LR method is a useful tool for estimating improvements in accuracy of EBVs due to the inclusion of genomic information when traditional validation methods as k-fold validation and predictive ability are not applicable.

Genomics to accelerate genetic improvement in tilapia

Selective breeding of tilapia populations started in the early 1990s and over the past three decades tilapia has become one of the most important farmed freshwater species, being produced in more than 125 countries around the globe. Although genome assemblies have been available since 2011, most of the tilapia industry still depends on classical selection techniques using mass spawning or pedigree information to select for growth traits with reported genetic gains of up to 20% per generation. The involvement of international breeding companies and research institutions has resulted in the rapid development and application of genomic resources in the last few years. GWAS and genomic selection are expected to contribute to uncovering the genetic variants involved in economically relevant traits and increasing the genetic gain in selective breeding programs, respectively. Developments over the next few years will probably focus on achieving a deep understanding of genetic architecture of complex traits, as well as accelerating genetic progress in the selection for growth-, quality- and robustness-related traits. Novel phenotyping technologies (i.e. phenomics), lower-cost whole-genome sequencing approaches, functional genomics and gene editing tools will be crucial in future developments for the improvement of tilapia aquaculture.

Single-Step Genomic Evaluations from Theory to Practice: Using SNP Chips and Sequence Data in BLUPF90

Single-step genomic evaluation became a standard procedure in livestock breeding, and the main reason is the ability to combine all pedigree, phenotypes, and genotypes available into one single evaluation, without the need of post-analysis processing. Therefore, the incorporation of data on genotyped and non-genotyped animals in this method is straightforward. Since 2009, two main implementations of single-step were proposed. One is called single-step genomic best linear unbiased prediction (ssGBLUP) and uses single nucleotide polymorphism (SNP) to construct the genomic relationship matrix; the other is the single-step Bayesian regression (ssBR), which is a marker effect model. Under the same assumptions, both models are equivalent. In this review, we focus solely on ssGBLUP. The implementation of ssGBLUP into the BLUPF90 software suite was done in 2009, and since then, several changes were made to make ssGBLUP flexible to any model, number of traits, number of phenotypes, and number of genotyped animals. Single-step GBLUP from the BLUPF90 software suite has been used for genomic evaluations worldwide. In this review, we will show theoretical developments and numerical examples of ssGBLUP using SNP data from regular chips to sequence data.

Harnessing genomics to fast-track genetic improvement in aquaculture

Aquaculture is the fastest-growing farmed food sector and will soon become the primary source of fish and shellfish for human diets. In contrast to crop and livestock production, aquaculture production is derived from numerous, exceptionally diverse species that are typically in the early stages of domestication. Genetic improvement of production traits via well-designed, managed breeding programmes has great potential to help meet the rising seafood demand driven by human population growth. Supported by continuous advances in sequencing and bioinformatics, genomics is increasingly being applied across the broad range of aquaculture species and at all stages of the domestication process to optimize selective breeding. In the future, combining genomic selection with biotechnological innovations, such as genome editing and surrogate broodstock technologies, may further expedite genetic improvement in aquaculture.

Assessment of genetic diversity and population structure in cultured Australian Pacific oysters

The recent development of Pacific oyster (Crassostrea gigas) SNP genotyping arrays has allowed detailed characterisation of genetic diversity and population structure within and between oyster populations. It also raises the potential of harnessing genomic selection for genetic improvement in oyster breeding programmes. The aim of this study was to characterise a breeding population of Australian oysters through genotyping and analysis of 18 027 SNPs, followed by comparison with genotypes of oyster sampled from Europe and Asia. This revealed that the Australian populations had similar population diversity (H_E ) to oysters from New Zealand, the British Isles, France and Japan. Population divergence was assessed using PCA of genetic distance and revealed that Australian oysters were distinct from all other populations tested. Australian Pacific oysters originate from planned introductions sourced from three Japanese populations. Approximately 95% of these introductions were from geographically, and potentially genetically, distinct populations from the Nagasaki oysters assessed in this study. Finally, in preparation for the application of genomic selection in oyster breeding programmes, the strength of LD was evaluated and subsets of loci were tested for their ability to accurately infer relationships. Weak LD was observed on average; however, SNP subsets were shown to accurately reconstitute a genomic relationship matrix constructed using all loci. This suggests that low-density SNP panels may have utility in the Australian population tested, and the findings represent an important first step towards the design and implementation of genomic approaches for applied breeding in Pacific oysters.

Genome-Wide Association Study and Cost-Efficient Genomic Predictions for Growth and Fillet Yield in Nile Tilapia (Oreochromis niloticus)

Fillet yield (FY) and harvest weight (HW) are economically important traits in Nile tilapia production. Genetic improvement of these traits, especially for FY, are lacking, due to the absence of efficient methods to measure the traits without sacrificing fish and the use of information from relatives to selection. However, genomic information could be used by genomic selection to improve traits that are difficult to measure directly in selection candidates, as in the case of FY. The objectives of this study were: (i) to perform genome-wide association studies (GWAS) to dissect the genetic architecture of FY and HW, (ii) to evaluate the accuracy of genotype imputation and (iii) to assess the accuracy of genomic selection using true and imputed low-density (LD) single nucleotide polymorphism (SNP) panels to determine a cost-effective strategy for practical implementation of genomic information in tilapia breeding programs. The data set consisted of 5,866 phenotyped animals and 1,238 genotyped animals (108 parents and 1,130 offspring) using a 50K SNP panel. The GWAS were performed using all genotyped and phenotyped animals. The genotyped imputation was performed from LD panels (LD0.5K, LD1K and LD3K) to high-density panel (HD), using information from parents and 20% of offspring in the reference set and the remaining 80% in the validation set. In addition, we tested the accuracy of genomic selection using true and imputed genotypes comparing the accuracy obtained from pedigree-based best linear unbiased prediction (PBLUP) and genomic predictions. The results from GWAS supports evidence of the polygenic nature of FY and HW. The accuracy of imputation ranged from 0.90 to 0.98 for LD0.5K and LD3K, respectively. The accuracy of genomic prediction outperformed the estimated breeding value from PBLUP. The use of imputation for genomic selection resulted in an increased relative accuracy independent of the trait and LD panel analyzed. The present results suggest that genotype imputation could be a cost-effective strategy for genomic selection in Nile tilapia breeding programs.

Genome-wide estimates of genetic diversity, inbreeding and effective size of experimental and commercial rainbow trout lines undergoing selective breeding

BACKGROUND

Selective breeding is a relatively recent practice in aquaculture species compared to terrestrial livestock. Nevertheless, the genetic variability of farmed salmonid lines, which have been selected for several generations, should be assessed. Indeed, a significant decrease in genetic variability due to high selection intensity could have occurred, potentially jeopardizing the long-term genetic progress as well as the adaptive capacities of populations facing change(s) in the environment. Thus, it is important to evaluate the impact of selection practices on genetic diversity to limit future inbreeding. The current study presents an analysis of genetic diversity within and between six French rainbow trout (Oncorhynchus mykiss) experimental or commercial lines based on a medium-density single nucleotide polymorphism (SNP) chip and various molecular genetic indicators: fixation index (FST), linkage disequilibrium (LD), effective population size (Ne) and inbreeding coefficient derived from runs of homozygosity (ROH).

RESULTS

Our results showed a moderate level of genetic differentiation between selected lines (FST ranging from 0.08 to 0.15). LD declined rapidly over the first 100 kb, but then remained quite high at long distances, leading to low estimates of Ne in the last generation ranging from 24 to 68 depending on the line and methodology considered. These results were consistent with inbreeding estimates that varied from 10.0% in an unselected experimental line to 19.5% in a commercial line, and which are clearly higher than corresponding estimates in ruminants or pigs. In addition, strong variations in LD and inbreeding were observed along the genome that may be due to differences in local rates of recombination or due to key genes that tended to have fixed favorable alleles for domestication or production.

CONCLUSIONS

This is the first report on ROH for any aquaculture species. Inbreeding appeared to be moderate to high in the six French rainbow trout lines, due to founder effects at the start of the breeding programs, but also likely to sweepstakes reproductive success in addition to selection for the selected lines. Efficient management of inbreeding is a major goal in breeding programs to ensure that populations can adapt to future breeding objectives and SNP information can be used to manage the rate at which inbreeding builds up in the fish genome.