Linear Unbiased Prediction Research Articles

Genomic prediction for sugarcane diseases including hybrid Bayesian-machine learning approaches.

Sugarcane smut and Pachymetra root rots are two serious diseases of sugarcane, with susceptible infected crops losing over 30% of yield. A heritable component to both diseases has been demonstrated, suggesting selection could improve disease resistance. Genomic selection could accelerate gains even further, enabling early selection of resistant seedlings for breeding and clonal propagation. In this study we evaluated four types of algorithms for genomic predictions of clonal performance for disease resistance. These algorithms were: Genomic best linear unbiased prediction (GBLUP), including extensions to model dominance and epistasis, Bayesian methods including BayesC and BayesR, Machine learning methods including random forest, multilayer perceptron (MLP), modified convolutional neural network (CNN) and attention networks designed to capture epistasis across the genome-wide markers. Simple hybrid methods, that first used BayesR/GWAS to identify a subset of 1000 markers with moderate to large marginal additive effects, then used attention networks to derive predictions from these effects and their interactions, were also developed and evaluated. The hypothesis for this approach was that using a subset of markers more likely to have an effect would enable better estimation of interaction effects than when there were an extremely large number of possible interactions, especially with our limited data set size. To evaluate the methods, we applied both random five-fold cross-validation and a structured PCA based cross-validation that separated 4702 sugarcane clones (that had disease phenotypes and genotyped for 26k genome wide SNP markers) by genomic relationship. The Bayesian methods (BayesR and BayesC) gave the highest accuracy of prediction, followed closely by hybrid methods with attention networks. The hybrid methods with attention networks gave the lowest variation in accuracy of prediction across validation folds (and lowest MSE), which may be a criteria worth considering in practical breeding programs. This suggests that hybrid methods incorporating the attention mechanism could be useful for genomic prediction of clonal performance, particularly where non-additive effects may be important.

Approximation of reliabilities for random-regression single-step genomic best linear unbiased predictor models

Random-regression models (RRM) are used in national genetic evaluations for longitudinal traits. The outputs of RRM are an index based on random regression coefficients and its reliability. The reliabilities are obtained from the inverse of the coefficient matrix of mixed model equations (MME). The reliabilities must be approximated for large data sets because it is impossible to invert the MME. There is no extensive literature on methods to approximate the reliabilities of RRM when genomic information is included by single-step GBLUP. We developed an algorithm to approximate such reliabilities. Our method combines the reliability of the index without genomic information with the reliability of a GBLUP model in terms of effective record contributions. We tested our algorithm in the 3-lactation model for milk yield from the Czech Republic. The data had 30 million test-day records, 2.5 million animals in the pedigree, and 54,000 genotyped animals. The correlation between our approximation and the reliabilities obtained from the inversion of the MME was 0.98, and the slope and intercept of the regression were 0.91 and 0.02, respectively. The elapsed time to approximate the reliabilities for the Czech data was 21 min.

Effect of genetic and sex effect on genomic prediction for horn development in Nellore cattle

This study aimed to evaluate the influence of phenotypic classification of horn development, animal sex effect, and non-autosomal SNP (single nucleotide polymorphism) markers on the genetic parameters and genomic prediction ability for horn development in Nellore cattle using the single-step genomic best linear unbiased prediction method. The polled phenotype was evaluated in two (presence and absence of horns), three (scurs and polled offspring from a horned parent, and the polled and horned animals), and four (absence of horn, polled born to a parent with horn, scurs, and presence of horn) phenotypic categories. A total of 12 statistical models were evaluated. The variance components were estimated using the THRGIBBS1F90 software, and a threshold animal model was used for genomic prediction analyses with the single-step genomic BLUP (ssGBLUP) procedure. Accuracy, bias, and dispersion parameters were evaluated based on the linear regression (LR) method. The highest heritability (0.84) was obtained when the polled character was evaluated as a binary trait. The lowest heritability estimates (0.44 to 0.45) for horn development were obtained when the phenotype was classified into three categories. For the same horn development classification method, the heritability estimates were similar regardless of the genomic evaluated models and fixed effects included in the model. For models considering four and three phenotypic categories for horn development, the inclusion of the sex effect as a fixed effect within the CG did not improve the accuracy, bias, and dispersion of genomic predictions for horn development. Analyzing the trait with binary expression, the highest prediction accuracy was observed when the effect of sex was not included in the CG and without the SNPs in the sex chromosomes. These models displayed the highest dispersion, pointing out the low robustness of genomic prediction. In addition, models that use less than four categories to classify the horn development phenotype, with no discrimination between polled and homozygous polled displayed lower prediction ability. The inclusion of non-autosomal SNPs in the analyses for the models considering four phenotypic categories leads to an improvement in prediction accuracy in 5,26 %, bias, and dispersion reduction, 37 % and 4,55 %, respectively, compared with models that only considered autosomal SNPs. The selection using genomic information for the polled trait is feasible, and it is an alternative to obtaining polled Nellore animals. The binary coding of horn development is an unsuitable oversimplification of polled phenotype, and probably, the genetic background of horn development is more complex than previously proposed. The most adequate prediction model to evaluate the horn development in Nellore cattle was considering four phenotypic categories and including non-autosomal SNP in the analyses for genomic prediction purposes of naturally genetically polled animals. Genetic dehorning can be adopted on a large scale as a low-cost and non-invasive approach to increase the frequency of hornless animals using genomic information and mating strategies.

Spatial best linear unbiased prediction: a computational mathematics approach for high dimensional massive datasets

Spatial best linear unbiased prediction: a computational mathematics approach for high dimensional massive datasets

Genomic prediction of blood biomarkers of metabolic disorders in Holstein cattle using parametric and nonparametric models

BackgroundMetabolic disturbances adversely impact productive and reproductive performance of dairy cattle due to changes in endocrine status and immune function, which increase the risk of disease. This may occur in the post-partum phase, but also throughout lactation, with sub-clinical symptoms. Recently, increased attention has been directed towards improved health and resilience in dairy cattle, and genomic selection (GS) could be a helpful tool for selecting animals that are more resilient to metabolic disturbances throughout lactation. Hence, we evaluated the genomic prediction of serum biomarkers levels for metabolic distress in 1353 Holsteins genotyped with the 100K single nucleotide polymorphism (SNP) chip assay. The GS was evaluated using parametric models best linear unbiased prediction (GBLUP), Bayesian B (BayesB), elastic net (ENET), and nonparametric models, gradient boosting machine (GBM) and stacking ensemble (Stack), which combines ENET and GBM approaches.ResultsThe results show that the Stack approach outperformed other methods with a relative difference (RD), calculated as an increment in prediction accuracy, of approximately 18.0% compared to GBLUP, 12.6% compared to BayesB, 8.7% compared to ENET, and 4.4% compared to GBM. The highest RD in prediction accuracy between other models with respect to GBLUP was observed for haptoglobin (hapto) from 17.7% for BayesB to 41.2% for Stack; for Zn from 9.8% (BayesB) to 29.3% (Stack); for ceruloplasmin (CuCp) from 9.3% (BayesB) to 27.9% (Stack); for ferric reducing antioxidant power (FRAP) from 8.0% (BayesB) to 40.0% (Stack); and for total protein (PROTt) from 5.7% (BayesB) to 22.9% (Stack). Using a subset of top SNPs (1.5k) selected from the GBM approach improved the accuracy for GBLUP from 1.8 to 76.5%. However, for the other models reductions in prediction accuracy of 4.8% for ENET (average of 10 traits), 5.9% for GBM (average of 21 traits), and 6.6% for Stack (average of 16 traits) were observed.ConclusionsOur results indicate that the Stack approach was more accurate in predicting metabolic disturbances than GBLUP, BayesB, ENET, and GBM and seemed to be competitive for predicting complex phenotypes with various degrees of mode of inheritance, i.e. additive and non-additive effects. Selecting markers based on GBM improved accuracy of GBLUP.

Genomic prediction based on a joint reference population for the Xinjiang Brown cattle.

Introduction: Xinjiang Brown cattle constitute the largest breed of cattle in Xinjiang. Therefore, it is crucial to establish a genomic evaluation system, especially for those with low levels of breed improvement. Methods: This study aimed to establish a cross breed joint reference population by analyzing the genetic structure of 485 Xinjiang Brown cattle and 2,633 Chinese Holstein cattle (Illumina GeneSeek GGP bovine 150K chip). The Bayes method single-step genome-wide best linear unbiased prediction was used to conduct a genomic evaluation of the joint reference population for the milk traits of Xinjiang Brown cattle. The reference population of Chinese Holstein cattle was randomly divided into groups to construct the joint reference population. By comparing the prediction accuracy, estimation bias, and inflation coefficient of the validation population, the optimal number of joint reference populations was determined. Results and Discussion: The results indicated a distinct genetic structure difference between the two breeds of adult cows, and both breeds should be considered when constructing multi-breed joint reference and validation populations. The reliability range of genome prediction of milk traits in the joint reference population was 0.142-0.465. Initially, it was determined that the inclusion of 600 and 900 Chinese Holstein cattle in the joint reference population positively impacted the genomic prediction of Xinjiang Brown cattle to certain extent. It was feasible to incorporate the Chinese Holstein into Xinjiang Brown cattle population to form a joint reference population for multi-breed genomic evaluation. However, for different Xinjiang Brown cattle populations, a fixed number of Chinese Holstein cattle cannot be directly added during multi-breed genomic selection. Pre-evaluation analysis based on the genetic structure, kinship, and other factors of the current population is required to ensure the authenticity and reliability of genomic predictions and improve estimation accuracy.

A combination of QTL mapping and genome-wide association study revealed the key gene for husk number in maize.

Two key genes Zm00001d021232 and Zm00001d048138 were identified by QTL mapping and GWAS. Additionally, they were verified to be significantly associated with maize husk number (HN) using gene-based association study. As a by-product of maize production, maize husk is an important industrial raw material. Husk layer number (HN) is an important trait that affects the yield of maize husk. However, the genetic mechanism underlying HN remains unclear. Herein, a total of 13 quantitative trait loci (QTL) controlling HN were identified in an IBM Syn 10 DH population across different locations. Among these, three QTL were individually repeatedly detected in at least two environments. Meanwhile, 26 unique single nucleotide polymorphisms (SNPs) were detected to be significantly (p < 2.15 × 10-6) associated with HN in an association pool. Of these SNPs, three were simultaneously detected across multiple environments or environments and best linear unbiased prediction (BLUP). We focused on these environment-stable and population-common genetic loci for excavating the candidate genes responsible for maize HN. Finally, 173 initial candidate genes were identified, of which 22 were involved in both multicellular organism development and single-multicellular organism process and thus confirmed as the candidate genes for HN. Gene-based association analyses revealed that the variants in four genes were significantly (p < 0.01/N) correlated with HN, of which Zm00001d021232 and Zm00001d048138 were highly expressed in husks and early developing ears among different maize tissues. Our study contributes to the understanding of genetic and molecular mechanisms of maize husk yield and industrial development in the future.

Marker assisted breeding of Sub1 introgressed rice (Oryza sativa L.) lines and identification of stable variety MTU 1232 suitable for flood prone ecosystem

The Development of flood-tolerant rice varieties is a prerequisite for climate resilience in flood-prone areas. The present study aimed to develop a stable, high, yielding, and tolerant rice variety against flash floods and stagnant flooding across multiple environments. Sub1A was incorporated into a popular rice variety MTU 1075 using Swarna-sub1 as a donor to generate BC3F5 families. Sub1BC2 was used as a foreground marker for selection, a proxy for the Sub1A gene. RM23865 and RM464 on Chromosome 9 were used as recombinant markers. Backcross families from the BC3F2 generation were evaluated under two weeks of flash floods 15 days after transplanting. This was followed by stagnant flooding and survived BC3F4 families were used for background selection using a 50K high-density SNP chip. The nine best families identified were included in the field trial evaluation under eight environments. Consequently, MTU Rice 1232 was identified as a high-yielding, flood-tolerant rice variety using Additive Main effects and Multiplicative Interaction stability analysis. The MTU Rice 1232 ranked first by the Best linear unbiased prediction (BLUP) based ranking parameters and 4th based on stability parameters ranking. This flood-tolerant rice variety can tolerate both flash floods and stagnant flooding and possesses an 80% survival rate. It has a yield potential of 3792 Kg ha-1 under severe floods and 6000 Kg ha-1 under normal conditions.

Identification of adaptable sunflower (Helianthus annuus L.) genotypes using yield performance and multiple-traits index

Sunflower is the most important oil crop ranked as fourth edible oil in the world. The study was conducted in Northern Ethiopia during 2017–2019 cropping seasons using randomized completely block design with three replications. The objective was to decipher the genotype by environment interaction (GEI) in multi-environment trials (MET) and identify adaptable sunflower genotypes. Combined ANOVA, AMMI ANOVA and Eberhart and Rusell regression were analyzed, and GGE bi-plots, AMMI1 and AMMI2 bi-plots, Principal component Analysis (PCA), multi-trait genotype–ideotype distance index (MGIDI), correlation network plot for sunflower traits were sketched. AMMI stability measures, Best Linear Unbiased Prediction (BLUP) based indexes; parametric and non-parametric statistics were computed using R-statistical software. In the AMMI ANOVA the main effects of the environment (E) (54.18 % SS), genotype (G) (16.9 % SS) and GEI (23.50 % SS) were significant (p < 0.001). The genotypic Likely-hood Ratio Test revealed significant for all traits. The AMMI bi-plot and the GGE bi-plots selected G10 and G2 as the most adaptable genotypes. CV, HMGV, RPGV, HMRPGV, Pi, GAI, KRS, S(3) and S(6) also identified G10 as the most stable genotype. Based on the MGIDI, G10 (MGIDI = 1.45) and G5 (MGIDI = 2.19) are selected and these genotypes are recommended for further cultivation in Tigray.

Evaluation of Efficiency of Weighted Average of Stability and Mean Performance Estimated by Linear Mixed Models for Identifying High‐Yielding Lentil Genotypes Adapted to Rainfed Regions

ABSTRACTA combination of the two methods of stability analysis, the additive main effect and multiplicative interaction (AMMI) and the best linear unbiased prediction (BLUP), based on weighted average of stability (WAASB) estimated by the linear mixed models (LMM) index identified the improved genotypes. In this study, 17 advanced genotypes of lentil were studied at two locations, Zanjan and Maragheh, during the two seasons. To examine the genotype × environment interaction, the AMMI and BLUP methods using the WAASB and weighted average of mean performance (WAASBY) index were combined to evaluate the performance stability of genotypes according to different experimental plots. Considering the significant genotype × environment interaction based on likelihood ratio test (LRT), data were analyzed by the BLUP method. The highest grain yield was detected for genotype 13, followed by genotypes 7, 11, 20, 5, 12, and 19 with higher productivity than the grand mean. To select genotypes according to yield and stability, the WAASBY index was defined by combining mean grain yield and stability. Considering 50:50 contributions for the two components, the grain yield and stability of the 13 genotypes were higher than the grand mean. The highest WAASBY was observed for genotypes 7, 20, and 12, which were determined as the best genotypes of lentil under agroecological conditions encountered in the current study regions.

Accuracy of Genomic prediction for fleece traits in Inner Mongolia Cashmere goats

The fleece traits are important economic traits of goats. With the reduction of sequencing and genotyping cost and the improvement of related technologies, genomic selection for goats has become possible. The research collect pedigree, phenotype and genotype information of 2299 Inner Mongolia Cashmere goats (IMCGs) individuals. We estimate fixed effects, and compare the estimates of variance components, heritability and genomic predictive ability of fleece traits in IMCGs when using the pedigree based Best Linear Unbiased Prediction (ABLUP), Genomic BLUP (GBLUP) or single-step GBLUP (ssGBLUP). The fleece traits considered are cashmere production (CP), cashmere diameter (CD), cashmere length (CL) and fiber length (FL). It was found that year of production, sex, herd and individual ages had highly significant effects on the four fleece traits (P < 0.01). All of these factors should be considered when the genetic parameters of fleece traits in IMCGs are evaluated. The heritabilities of FL, CL, CP and CD with ABLUP, GBLUP and ssGBLUP methods were 0.26 ~ 0.31, 0.05 ~ 0.08, 0.15 ~ 0.20 and 0.22 ~ 0.28, respectively. Therefore, it can be inferred that the genetic progress of CL is relatively slow. The predictive ability of fleece traits in IMCGs with GBLUP (56.18% to 69.06%) and ssGBLUP methods (66.82% to 73.70%) was significantly higher than that of ABLUP (36.73% to 41.25%). For the ssGBLUP method is significantly (29% ~ 33%) higher than that with ABLUP, and which is slightly (4% ~ 14%) higher than that of GBLUP. The ssGBLUP will be as an superiors method for using genomic selection of fleece traits in Inner Mongolia Cashmere goats.

BLUP Estimation and Genotype Stability in Arachis hypogaea L. Variety Testing using Mixed Model Equations

Background: One of the major goals of plant breeding is the selection of high yielding superior cultivars having wide or specific adaptation. However, there is a fluctuation in the annual production due to the sensitive behaviour of the genotypes under different environmental conditions referred to as Genotype by Environment Interaction (GEI). The current study aimed to study the contribution of GEI for the adaptation of groundnut lines for spring and/or kharif season. Methods: To assess the contribution of GEI, Multi-Environment Trials (METs) were conducted for 40 confectionery purpose groundnut genotypes at F9 generation along with checks, across three locations for two seasons (spring and kharif). The contribution of environmental effects, genotypic values and genotype × environment interaction values were obtained from genotypic variance-covariance matrix Gi = Σg⨂A using mixed models (MM) in Best linear unbiased predictions (BLUPs).The pooled data was first partitioned into fixed effects of sites across the seasons and BLUP genotypic values (Ggge). The BLUP genotypic values are further partitioned into genetic value (Gg) and their interaction with the environment (Gge) for the adaptability of genotypes across seasons. Result: The results of MET revealed the presence of significant crossover interaction. The demarcation of advance breeding lines for adaptability across the environment as well as for season specific adaptation was done for variety testing. Genotypes having moderate to high Gge values along with high Gg values in spring than kharif, owing to their better performance during the spring season. CGL-11, CGL-23 and CGL-04 were the highest yielding genotypes, with quite high Gge values. This is due to the more favourable environmental conditions interacting positively with genotypes during the spring. Thus, the high Gg value(s) of genotype(s) alone is not a capable factor for commercialization as Gge value is the deciding factor for the adaptability for the targeted season.

Heritability and predictive ability for the heterobothriosis resistance and growth performance in the tiger pufferfish Takifugu rubripes fed standard or low fishmeal diets

Heterobothriosis resistance and high growth performance are major concerns in the tiger pufferfish (Takifugu rubripes) farming. Our previous small-scale study suggested the feasibility of genomic selection (GS) for these traits but also identified an undesired genetic correlation between them, which could hamper simultaneous genetic improvement. Here, we sought to validate these initial findings through larger samples and crosses and also explore the potential of GS for the individuals fed a normal fishmeal diet and a low fishmeal (LFM) diet. We produced three populations: one was fed a normal diet (PopFM: 1100 individuals) and two were fed an LFM diet (PopLFM1: 993 individuals; PopLFM2: 1017 individuals); all three populations were artificially infected with Heterobothrium okamotoi, a monogenean parasite causing heterobothriosis. We recorded standard lengths of the fish after artificial infection (SLEAI) and the number of parasites on the gill cavity wall (HC) and genotyped genome-wide single nucleotide polymorphisms (SNPs) using genotyping by random amplicon sequencing, direct (GRAS-Di) technology for each sample: PopFM, 12 K SNPs; PopLFM1, 19 K SNPs; PopLFM2, 21 K SNPs. Heritability and prediction ability for each trait were estimated for each population using genomic best linear unbiased prediction (GBLUP), Bayes C, and Bayesian ridge regression methods. SNP effects on each trait were estimated by Bayes C, Bayesian ridge regression, and genome-wide association analyses (GWAS) using the same data sets. We observed moderate to high levels of heritability for HC (0.16–0.57) and SLEAI (0.50–0.77), with predictive ability of 0.25–0.55 for HC and of 0.51–0.73 for SLEAI. These values were slightly larger in LFM diet groups. As reported previously, a weak undesired genetic correlation was observed between the traits, and the correlation was more pronounced in the LFM groups. In conclusion, our results indicate that genetic gain in growth and parasite resistance traits can be expected for the tiger pufferfish through GS, especially for individuals reared on an LFM diet. However, close attention should be paid to the extent of the undesirable genetic correlation between the two traits that may be amplified in the populations fed an LFM diet.

Genomic Prediction of Growth Traits in Yorkshire Pigs of Different Reference Group Sizes Using Different Estimated Breeding Value Models.

The need for sufficient reference population data poses a significant challenge in breeding programs aimed at improving pig farming on a small to medium scale. To overcome this hurdle, investigating the advantages of combing reference populations of varying sizes is crucial for enhancing the accuracy of the genomic estimated breeding value (GEBV). Genomic selection (GS) in populations with limited reference data can be optimized by combining populations of the same breed or related breeds. This study focused on understanding the effect of combing different reference group sizes on the accuracy of GS for determining the growth effectiveness and percentage of lean meat in Yorkshire pigs. Specifically, our study investigated two important traits: the age at 100 kg live weight (AGE100) and the backfat thickness at 100 kg live weight (BF100). This research assessed the efficiency of genomic prediction (GP) using different GEBV models across three Yorkshire populations with varying genetic backgrounds. The GeneSeek 50K GGP porcine high-density array was used for genotyping. A total of 2295 Yorkshire pigs were included, representing three Yorkshire pig populations with different genetic backgrounds-295 from Danish (small) lines from Huaibei City, Anhui Province, 500 from Canadian (medium) lines from Lixin County, Anhui Province, and 1500 from American (large) lines from Shanghai. To evaluate the impact of different population combination scenarios on the GS accuracy, three approaches were explored: (1) combining all three populations for prediction, (2) combining two populations to predict the third, and (3) predicting each population independently. Five GEBV models, including three Bayesian models (BayesA, BayesB, and BayesC), the genomic best linear unbiased prediction (GBLUP) model, and single-step GBLUP (ssGBLUP) were implemented through 20 repetitions of five-fold cross-validation (CV). The results indicate that predicting one target population using the other two populations yielded the highest accuracy, providing a novel approach for improving the genomic selection accuracy in Yorkshire pigs. In this study, it was found that using different populations of the same breed to predict small- and medium-sized herds might be effective in improving the GEBV. This investigation highlights the significance of incorporating population combinations in genetic models for predicting the breeding value, particularly for pig farmers confronted with resource limitations.

Tomato families possessing resistance to late blight also display high-quality fruit

In recent years, several efforts have been made to develop tomato cultivars displaying both late blight resistance and good organoleptic fruit quality. Selection indexes are considered the best option to perform genotype selection when many different traits are being considered to select genotypes as close to the desired ideotype as possible. Therefore, this study aimed at selecting late blight-resistant tomato families based on their fruit quality attributes using factor analysis and ideotype-design / best linear unbiased predictor (FAI-BLUP) index. For this purpose, we assessed the fruit quality parameters of 81 F3:5 tomato families previously selected as late blight resistant. The tomato cultivars Thaise, Argos, and Liberty were included in the trial as checks. The experimental arrangement consisted of complete randomized blocks with three replicates. Each plot was formed by five plants, three of which were used in the fruit quality assessment. The quality parameters assessed were fruit diameter, fruit length, fruit color (L, a*, C, and H), fruit firmness, titratable acidity, soluble solids content, hydrogen potential, and SS:TA ratio. Fruit quality data were analyzed using the mixed model methodology via REML/BLUP (restricted residual maximum likelihood / best linear unbiased prediction) to obtain BLUPs that were further subjected to the FAI-BLUP selection index. The FAI-BLUP was efficient in selecting late blight-resistant tomato genotypes based on their fruit quality attributes. Fourteen tomato families were classified as closest to the desirable ideotype for fruit quality. These genotypes should move on to the following stages of the tomato breeding program.

Marker Density and Models to Improve the Accuracy of Genomic Selection for Growth and Slaughter Traits in Meat Rabbits.

The selection and breeding of good meat rabbit breeds are fundamental to their industrial development, and genomic selection (GS) can employ genomic information to make up for the shortcomings of traditional phenotype-based breeding methods. For the practical implementation of GS in meat rabbit breeding, it is necessary to assess different marker densities and GS models. Here, we obtained low-coverage whole-genome sequencing (lcWGS) data from 1515 meat rabbits (including parent herd and half-sibling offspring). The specific objectives were (1) to derive a baseline for heritability estimates and genomic predictions based on randomly selected marker densities and (2) to assess the accuracy of genomic predictions for single- and multiple-trait linear mixed models. We found that a marker density of 50 K can be used as a baseline for heritability estimation and genomic prediction. For GS, the multi-trait genomic best linear unbiased prediction (GBLUP) model results in more accurate predictions for virtually all traits compared to the single-trait model, with improvements greater than 15% for all of them, which may be attributed to the use of information on genetically related traits. In addition, we discovered a positive correlation between the performance of the multi-trait GBLUP and the genetic correlation between the traits. We anticipate that this approach will provide solutions for GS, as well as optimize breeding programs, in meat rabbits.

Application of deep learning with bivariate models for genomic prediction of sow lifetime productivity-related traits.

Pig breeders cannot obtain phenotypic information at the time of selection for sow lifetime productivity (SLP). They would benefit from obtaining genetic information of candidate sows. Genomic data interpreted using deep learning (DL) techniques could contribute to the genetic improvement of SLP to maximize farm profitability because DL models capture nonlinear genetic effects such as dominance and epistasis more efficiently than conventional genomic prediction methods based on linear models. This study aimed to investigate the usefulness of DL for the genomic prediction of two SLP-related traits; lifetime number of litters (LNL) and lifetime pig production (LPP). Two bivariate DL models, convolutional neural network (CNN) and local convolutional neural network (LCNN), were compared with conventional bivariate linear models (i.e., genomic best linear unbiased prediction, Bayesian ridge regression, Bayes A, and Bayes B). Phenotype and pedigree data were collected from 40,011 sows that had husbandry records. Among these, 3,652 pigs were genotyped using the PorcineSNP60K BeadChip. The best predictive correlation for LNL was obtained with CNN (0.28), followed by LCNN (0.26) and conventional linear models (approximately 0.21). For LPP, the best predictive correlation was also obtained with CNN (0.29), followed by LCNN (0.27) and conventional linear models (approximately 0.25). A similar trend was observed with the mean squared error of prediction for the SLP traits. This study provides an example of a CNN that can outperform against the linear model-based genomic prediction approaches when the nonlinear interaction components are important because LNL and LPP exhibited strong epistatic interaction components. Additionally, our results suggest that applying bivariate DL models could also contribute to the prediction accuracy by utilizing the genetic correlation between LNL and LPP.

Selection of informative markers using machine learning approaches and genome-wide association studies to improve genomic prediction in Hanwoo cattle: a simulation study

The present study deploys a comparison of Gradient Boosting Machine (GBM), Extreme Gradient Boosting (XGBoost), and Genome Wide Association Studies (GWAS) in selecting optimum subsets of single nucleotide polymorphisms (SNPs) to be used in genomic prediction in cattle. The data simulation was carried out for 6,000 animals and 47,841 SNPs which include 43,633 polygenic markers and 4208 quantitative trait loci (QTL) using QMSim software. The genomic prediction was conducted with the best linear unbiased prediction (BLUP) method using the BLUPF90 program. The accuracy of prediction was computed in three different types, namely, Empirical all SNPs, Empirical QTL, and theoretical accuracy, Accuracy PEV . Among the three models, the highest Empirical all SNPs accuracy 0.79 was derived for GBM followed by 0.77 for XGBoost and 0.76 for GWAS. The Empirical QTL accuracy was almost equal for all three models. The maximum theoretical accuracy was obtained for GWAS which was 0.93, whereas GBM and XGBoost obtained 0.86 and 0.85 accuracy levels respectively. Our results indicate that all three models comparably performed in genomic predictions; however, subsets selected by both GBM and GWAS reported higher prediction accuracies compared to the whole SNP set. The number of QTL selected as a proportion of the total number of SNPs was superior in GWAS. These observations can be validated using real data which could enable further optimization of the analysis process.

Comparative analysis of the heritability, estimated breeding value, and accuracy of carcass traits in Hanwoo from the Yeongnam Region, considering breeding area and climate effects, utilizing genomic information

Climate change is causing abnormal weather patterns worldwide. In Korea, local governments are characterized by various climatic conditions. Therefore, research integrating climate factors with Hanwoo is necessary. In this study, 562 Hanwoo were analyzed using Pedigree Best Linear Unbiased Prediction (PBLUP) and Genomic BLUP (GBLUP) models while considering climate factors such as temperature and humidity in the Yeongnam region. Carcass Weight (CWT) varied significantly in PBLUP depending on the model, while the heritability of Eye Muscle Area (EMA) and Back-Fat Thickness (BFT) decreased. However, the Marbling Score (MS) maintained high heritability across all models. GBLUP showed a slight decrease in the analysis model that considers breeding areas and climatic factors, yet still led to a high estimate of heritability for all traits. However, MS consistently had a heritability of 0.50 across all models. In PBLUP, the accuracy of the Estimated Breeding Value (EBV) increased by 0.15 for CWT (0.62-0.77) compared to Model 1. However, a decrease of 0.01–0.06 was observed for EMA (0.62–0.68) and BFT (0.53–0.58), while an increase of 0.01 was noted for MS (0.72–0.73). The accuracy of EBV using GBLUP was as follows: CWT (0.75-0.77), EMA (0.72-0.80), BFT (0.71-0.81), and MS (0.75). The accuracy of BFT significantly increased compared to PBLUP. For MS, it did not appear to be influenced by breeding area and climatic conditions. These results suggest the need for ongoing research on the relationship between Hanwoo carcass traits and climate effects, requiring additional test groups from various breeding areas.

Flagging unusual clusters based on linear mixed models using weighted and self-calibrated predictors.

Statistical models incorporating cluster-specific intercepts are commonly used in hierarchical settings, for example, observations clustered within patients or patients clustered within hospitals. Predicted values of these intercepts are often used to identify or "flag" extreme or outlying clusters, such as poorly performing hospitals or patients with rapid declines in their health. We consider a variety of flagging rules, assessing different predictors, and using different accuracy measures. Using theoretical calculations and comprehensive numerical evaluation, we show that previously proposed rules based on the 2 most commonly used predictors, the usual best linear unbiased predictor and fixed effects predictor, perform extremely poorly: the incorrect flagging rates are either unacceptably high (approaching 0.5 in the limit) or overly conservative (eg, much <0.05 for reasonable parameter values, leading to very low correct flagging rates). We develop novel methods for flagging extreme clusters that can control the incorrect flagging rates, including very simple-to-use versions that we call "self-calibrated." The new methods have substantially higher correct flagging rates than previously proposed methods for flagging extreme values, while controlling the incorrect flagging rates. We illustrate their application using data on length of stay in pediatric hospitals for children admitted for asthma diagnoses.