Recent advances in artificial intelligence, mechanistic models, and speed breeding offer exciting opportunities for precise and accelerated genomics-assisted breeding

Given the challenges of population growth and climate change, there is an urgent need to expedite the development of high-yielding stress-tolerant crop cultivars. While traditional breeding methods have been instrumental in ensuring global food security, their efficiency, precision, and labour intensiveness have become increasingly inadequate to address present and future challenges. Fortunately, recent advances in high-throughput phenomics and genomics-assisted breeding (GAB) provide a promising platform for enhancing crop cultivars with greater efficiency. However, several obstacles must be overcome to optimize the use of these techniques in crop improvement, such as the complexity of phenotypic analysis of big image data. In addition, the prevalent use of linear models in genome-wide association studies (GWAS) and genomic selection (GS) fails to capture the nonlinear interactions of complex traits, limiting their applicability for GAB and impeding crop improvement. Recent advances in artificial intelligence (AI) techniques have opened doors to nonlinear modelling approaches in crop breeding, enabling the capture of nonlinear and epistatic interactions in GWAS and GS and thus making this variation available for GAB. While statistical and software challenges persist in AI-based models, they are expected to be resolved soon. Furthermore, recent advances in speed breeding have significantly reduced the time (3-5-fold) required for conventional breeding. Thus, integrating speed breeding with AI and GAB could improve crop cultivar development within a considerably shorter timeframe while ensuring greater accuracy and efficiency. In conclusion, this integrated approach could revolutionize crop breeding paradigms and safeguard food production in the face of population growth and climate change.

Genetic architecture of inter-specific and -generic grass hybrids by network analysis on multi-omics data

BACKGROUND

Understanding the mechanisms underlining forage production and its biomass nutritive quality at the omics level is crucial for boosting the output of high-quality dry matter per unit of land. Despite the advent of multiple omics integration for the study of biological systems in major crops, investigations on forage species are still scarce.

RESULTS

Our results identified substantial changes in gene co-expression and metabolite-metabolite network topologies as a result of genetic perturbation by hybridizing L. perenne with another species within the genus (L. multiflorum) relative to across genera (F. pratensis). However, conserved hub genes and hub metabolomic features were detected between pedigree classes, some of which were highly heritable and displayed one or more significant edges with agronomic traits in a weighted omics-phenotype network. In spite of tagging relevant biological molecules as, for example, the light-induced rice 1 (LIR1), hub features were not necessarily better explanatory variables for omics-assisted prediction than features stochastically sampled and all available regressors.

CONCLUSIONS

The utilization of computational techniques for the reconstruction of co-expression networks facilitates the identification of key omic features that serve as central nodes and demonstrate correlation with the manifestation of observed traits. Our results also indicate a robust association between early multi-omic traits measured in a greenhouse setting and phenotypic traits evaluated under field conditions.

Machine learning in TCM with natural products and molecules: current status and future perspectives

Traditional Chinese medicine (TCM) has been practiced for thousands of years with clinical efficacy. Natural products and their effective agents such as artemisinin and paclitaxel have saved millions of lives worldwide. Artificial intelligence is being increasingly deployed in TCM. By summarizing the principles and processes of deep learning and traditional machine learning algorithms, analyzing the application of machine learning in TCM, reviewing the results of previous studies, this study proposed a promising future perspective based on the combination of machine learning, TCM theory, chemical compositions of natural products, and computational simulations based on molecules and chemical compositions. In the first place, machine learning will be utilized in the effective chemical components of natural products to target the pathological molecules of the disease which could achieve the purpose of screening the natural products on the basis of the pathological mechanisms they target. In this approach, computational simulations will be used for processing the data for effective chemical components, generating datasets for analyzing features. In the next step, machine learning will be used to analyze the datasets on the basis of TCM theories such as the superposition of syndrome elements. Finally, interdisciplinary natural product-syndrome research will be established by unifying the results of the two steps outlined above, potentially realizing an intelligent artificial intelligence diagnosis and treatment model based on the effective chemical components of natural products under the guidance of TCM theory. This perspective outlines an innovative application of machine learning in the clinical practice of TCM based on the investigation of chemical molecules under the guidance of TCM theory.

GWAS and genetic and phenotypic correlations of plasma metabolites with complete blood count traits in healthy young pigs reveal implications for pig immune response

Introduction: In this study estimated genetic and phenotypic correlations between fifteen complete blood count (CBC) traits and thirty-three heritable plasma metabolites in young healthy nursery pigs. In addition, it provided an opportunity to identify candidate genes associated with variation in metabolite concentration and their potential association with immune response, disease resilience, and production traits.

Methods: The blood samples were collected from healthy young pigs and Nuclear Magnetic Resonance (NMR) was used to quantify plasma metabolites. CBC was determined using the ADVIA® 2120i Hematology System. Genetic correlations of metabolite with CBC traits and single step genome-wide association study (ssGWAS) were estimated using the BLUPF90 programs.

Results: Results showed low phenotypic correlation estimates between plasma metabolites and CBC traits. The highest phenotypic correlation was observed between lactic acid and plasma basophil concentration (0.36 ± 0.04; p < 0.05). Several significant genetic correlations were found between metabolites and CBC traits. The plasma concentration of proline was genetically positively correlated with hemoglobin concentration (0.94 ± 0.03; p < 0.05) and L-tyrosine was negatively correlated with mean corpuscular hemoglobin (MCH; −0.92 ± 0.74; p < 0.05). The genomic regions identified in this study only explained a small percentage of the genetic variance of metabolites levels that were genetically correlated with CBC, resilience, and production traits.

Discussion: The results of this systems approach suggest that several plasma metabolite phenotypes are phenotypically and genetically correlated with CBC traits, suggesting that they may be potential genetic indicators of immune response following disease challenge. Genomic analysis revealed genes and pathways that might interact to modulate CBC, resilience, and production traits.

Re-Evaluation of Genotyping Methodologies in Cattle: The Proficiency of Imputation

In dairy cattle, identifying polymorphisms that contribute to complex economical traits such as residual feed intake (RFI) is challenging and demands accurate genotyping. In this study, we compared imputed genotypes (n = 192 cows) to those obtained using the TaqMan and high-resolution melting (HRM) methods (n = 114 cows), for mutations in the FABP4 gene that had been suggested to have a large effect on RFI. Combining the whole genome sequence (n = 19 bulls) and the cows’ BovineHD BeadChip allowed imputing genotypes for these mutations that were verified by Sanger sequencing, whereas, an error rate of 11.6% and 10.7% were encountered for HRM and TaqMan, respectively. We show that this error rate seriously affected the linkage-disequilibrium analysis that supported this gene candidacy over other BTA14 gene candidates. Thus, imputation produced superior genotypes and should also be regarded as a method of choice to validate the reliability of the genotypes obtained by other methodologies that are prone to genotyping errors due to technical conditions. These results support the view that RFI is a complex trait and that searching for the causative sequence variation underlying cattle RFI should await the development of statistical methods suitable to handle additive and epistatic interactions.

Impact of epistasis effects on the accuracy of predicting phenotypic values of residual feed intake in U. S Holstein cows

The impact of genomic epistasis effects on the accuracy of predicting the phenotypic values of residual feed intake (RFI) in U.S. Holstein cows was evaluated using 6215 Holstein cows and 78,964 SNPs. Two SNP models and seven epistasis models were initially evaluated. Heritability estimates and the accuracy of predicting the RFI phenotypic values from 10-fold cross-validation studies identified the model with SNP additive effects and additive × additive (A×A) epistasis effects (A + A×A model) to be the best prediction model. Under the A + A×A model, additive heritability was 0.141, and A×A heritability was 0.263 that consisted of 0.260 inter-chromosome A×A heritability and 0.003 intra-chromosome A×A heritability, showing that inter-chromosome A×A effects were responsible for the accuracy increases due to A×A. Under the SNP additive model (A-only model), the additive heritability was 0.171. In the 10 validation populations, the average accuracy for predicting the RFI phenotypic values was 0.246 (with range 0.197-0.333) under A + A×A model and was 0.231 (with range of 0.188-0.319) under the A-only model. The average increase in the accuracy of predicting the RFI phenotypic values by the A + A×A model over the A-only model was 6.49% (with range of 3.02-14.29%). Results in this study showed A×A epistasis effects had a positive impact on the accuracy of predicting the RFI phenotypic values when combined with additive effects in the prediction model.

Epi-MEIF: detecting higher order epistatic interactions for complex traits using mixed effect conditional inference forests

Understanding the relationship between genetic variations and variations in complex and quantitative phenotypes remains an ongoing challenge. While Genome-wide association studies (GWAS) have become a vital tool for identifying single-locus associations, we lack methods for identifying epistatic interactions. In this article, we propose a novel method for higher-order epistasis detection using mixed effect conditional inference forest (epiMEIF). The proposed method is fitted on a group of single nucleotide polymorphisms (SNPs) potentially associated with the phenotype and the tree structure in the forest facilitates the identification of n-way interactions between the SNPs. Additional testing strategies further improve the robustness of the method. We demonstrate its ability to detect true n-way interactions via extensive simulations in both cross-sectional and longitudinal synthetic datasets. This is further illustrated in an application to reveal epistatic interactions from natural variations of cardiac traits in flies (Drosophila). Overall, the method provides a generalized way to identify higher-order interactions from any GWAS data, thereby greatly improving the detection of the genetic architecture underlying complex phenotypes.

A Random Forest-Based Genome-Wide Scan Reveals Fertility-Related Candidate Genes and Potential Inter-Chromosomal Epistatic Regions Associated With Age at First Calving in Nellore Cattle

This study aimed to perform a genome-wide association analysis (GWAS) using the Random Forest (RF) approach for scanning candidate genes for age at first calving (AFC) in Nellore cattle. Additionally, potential epistatic effects were investigated using linear mixed models with pairwise interactions between all markers with high importance scores within the tree ensemble non-linear structure. Data from Nellore cattle were used, including records of animals born between 1984 and 2015 and raised in commercial herds located in different regions of Brazil. The estimated breeding values (EBV) were computed and used as the response variable in the genomic analyses. After quality control, the remaining number of animals and SNPs considered were 3,174 and 360,130, respectively. Five independent RF analyses were carried out, considering different initialization seeds. The importance score of each SNP was averaged across the independent RF analyses to rank the markers according to their predictive relevance. A total of 117 SNPs associated with AFC were identified, which spanned 10 autosomes (2, 3, 5, 10, 11, 17, 18, 21, 24, and 25). In total, 23 non-overlapping genomic regions embedded 262 candidate genes for AFC. Enrichment analysis and previous evidence in the literature revealed that many candidate genes annotated close to the lead SNPs have key roles in fertility, including embryo pre-implantation and development, embryonic viability, male germinal cell maturation, and pheromone recognition. Furthermore, some genomic regions previously associated with fertility and growth traits in Nellore cattle were also detected in the present study, reinforcing the effectiveness of RF for pre-screening candidate regions associated with complex traits. Complementary analyses revealed that many SNPs top-ranked in the RF-based GWAS did not present a strong marginal linear effect but are potentially involved in epistatic hotspots between genomic regions in different autosomes, remarkably in the BTAs 3, 5, 11, and 21. The reported results are expected to enhance the understanding of genetic mechanisms involved in the biological regulation of AFC in this cattle breed.

Methodology for the identification of relevant loci for milk traits in dairy cattle, using machine learning algorithms

Machine learning methods were considered efficient in identifying single nucleotide polymorphisms (SNP) underlying a trait of interest. This study aimed to construct predictive models using machine learning algorithms, to identify loci that best explain the variance in milk traits of dairy cattle. Further objectives involved validating the results by comparison with reported relevant regions and retrieving the pathways overrepresented by the genes flanking relevant SNPs. Regression models using XGBoost (XGB), LightGBM (LGB), and Random Forest (RF) algorithms were trained using estimated breeding values for milk production (EBV_M), milk fat content (EBV_F) and milk protein content (EBV_P) as phenotypes and genotypes on 40417 SNPs as predictor variables. To evaluate their efficiency, metrics for actual vs. predicted values were determined in validation folds (XGB and LGB) and out-of-bag data (RF). Less than 4500 relevant SNPs were retrieved for each trait. Among the genes flanking them, signaling and transmembrane transporter activities were overrepresented.

The models trained:•

Predicted breeding values for animals not included in the dataset.

•

Were efficient in identifying a subset of SNPs explaining phenotypic variation.

The results obtained using XGB and LGB algorithms agreed with previous results. Therefore, the method proposed could be applied for future association studies on milk traits.

Genome-Enabled Prediction Methods Based on Machine Learning

Growth of artificial intelligence and machine learning (ML) methodology has been explosive in recent years. In this class of procedures, computers get knowledge from sets of experiences and provide forecasts or classification. In genome-wide based prediction (GWP), many ML studies have been carried out. This chapter provides a description of main semiparametric and nonparametric algorithms used in GWP in animals and plants. Thirty-four ML comparative studies conducted in the last decade were used to develop a meta-analysis through a Thurstonian model, to evaluate algorithms with the best predictive qualities. It was found that some kernel, Bayesian, and ensemble methods displayed greater robustness and predictive ability. However, the type of study and data distribution must be considered in order to choose the most appropriate model for a given problem.

Integrated Environmental and Genomic Analysis Reveals the Drivers of Local Adaptation in African Indigenous Chickens

Breeding for climate resilience is currently an important goal for sustainable livestock production. Local adaptations exhibited by indigenous livestock allow investigating the genetic control of this resilience. Ecological niche modeling (ENM) provides a powerful avenue to identify the main environmental drivers of selection. Here, we applied an integrative approach combining ENM with genome-wide selection signature analyses (XPEHH and Fst) and genotype-environment association (redundancy analysis), with the aim of identifying the genomic signatures of adaptation in African village chickens. By dissecting 34 agro-climatic variables from the ecosystems of 25 Ethiopian village chicken populations, ENM identified six key drivers of environmental challenges: One temperature variable-strongly correlated with elevation, three precipitation variables as proxies for water availability, and two soil/land cover variables as proxies of food availability for foraging chickens. Genome analyses based on whole-genome sequencing (n = 245), identified a few strongly supported genomic regions under selection for environmental challenges related to altitude, temperature, water scarcity, and food availability. These regions harbor several gene clusters including regulatory genes, suggesting a predominantly oligogenic control of environmental adaptation. Few candidate genes detected in relation to heat-stress, indicates likely epigenetic regulation of thermo-tolerance for a domestic species originating from a tropical Asian wild ancestor. These results provide possible explanations for the rapid past adaptation of chickens to diverse African agro-ecologies, while also representing new landmarks for sustainable breeding improvement for climate resilience. We show that the pre-identification of key environmental drivers, followed by genomic investigation, provides a powerful new approach for elucidating adaptation in domestic animals.

A Large-Scale Genome-Wide Association Study of Epistasis Effects of Production Traits and Daughter Pregnancy Rate in U.S. Holstein Cattle

Epistasis is widely considered important, but epistasis studies lag those of SNP effects. Genome-wide association study (GWAS) using 76,109 SNPs and 294,079 first-lactation Holstein cows was conducted for testing pairwise epistasis effects of five production traits and three fertility traits: milk yield (MY), fat yield (FY), protein yield (PY), fat percentage (FPC), protein percentage (PPC), and daughter pregnancy rate (DPR). Among the top 50,000 pairwise epistasis effects of each trait, the five production traits had large chromosome regions with intra-chromosome epistasis. The percentage of inter-chromosome epistasis effects was 1.9% for FPC, 1.6% for PPC, 10.6% for MY, 29.9% for FY, 39.3% for PY, and 84.2% for DPR. Of the 50,000 epistasis effects, the number of significant effects with log10(1/p) ≥ 12 was 50,000 for FPC and PPC, and 10,508, 4763, 4637 and 1 for MY, FY, PY and DPR, respectively, and A × A effects were the most frequent epistasis effects for all traits. Majority of the inter-chromosome epistasis effects of FPC across all chromosomes involved a Chr14 region containing DGAT1, indicating a potential regulatory role of this Chr14 region affecting all chromosomes for FPC. The epistasis results provided new understanding about the genetic mechanism underlying quantitative traits in Holstein cattle.

Joint contributions of the gut microbiota and host genetics to feed efficiency in chickens

BACKGROUND

Feed contributes most to livestock production costs. Improving feed efficiency is crucial to increase profitability and sustainability for animal production. Host genetics and the gut microbiota can both influence the host phenotype. However, the association between the gut microbiota and host genetics and their joint contribution to feed efficiency in chickens is largely unclear.

RESULTS

Here, we examined microbial data from the duodenum, jejunum, ileum, cecum, and feces in 206 chickens and their host genotypes and confirmed that the microbial phenotypes and co-occurrence networks exhibited dramatic spatial heterogeneity along the digestive tract. The correlations between host genetic kinship and gut microbial similarities within different sampling sites were weak, with coefficients ranging from - 0.07 to 0.08. However, microbial genome-wide analysis revealed that genetic markers near or inside the genes MTHFD1L and LARGE1 were associated with the abundances of cecal Megasphaera and Parabacteroides, respectively. The effect of host genetics on residual feed intake (RFI) was 39%. We further identified three independent genetic variations that were related to feed efficiency and had a modest effect on the gut microbiota. The contributions of the gut microbiota from the different parts of the intestinal tract on RFI were distinct. The cecal microbiota accounted for 28% of the RFI variance, a value higher than that explained by the duodenal, jejunal, ileal, and fecal microbiota. Additionally, six bacteria exhibited significant associations with RFI. Specifically, lower abundances of duodenal Akkermansia muciniphila and cecal Parabacteroides and higher abundances of cecal Lactobacillus, Corynebacterium, Coprobacillus, and Slackia were related to better feed efficiency.

CONCLUSIONS

Our findings solidified the notion that both host genetics and the gut microbiota, especially the cecal microbiota, can drive the variation in feed efficiency. Although host genetics has a limited effect on the entire microbial community, a small fraction of gut microorganisms tends to interact with host genes, jointly contributing to feed efficiency. Therefore, the gut microbiota and host genetic variations can be simultaneously targeted by favoring more-efficient taxa and selective breeding to improve feed efficiency in chickens. Video abstract.

Identification of Predictor Genes for Feed Efficiency in Beef Cattle by Applying Machine Learning Methods to Multi-Tissue Transcriptome Data

Machine learning (ML) methods have shown promising results in identifying genes when applied to large transcriptome datasets. However, no attempt has been made to compare the performance of combining different ML methods together in the prediction of high feed efficiency (HFE) and low feed efficiency (LFE) animals. In this study, using RNA sequencing data of five tissues (adrenal gland, hypothalamus, liver, skeletal muscle, and pituitary) from nine HFE and nine LFE Nellore bulls, we evaluated the prediction accuracies of five analytical methods in classifying FE animals. These included two conventional methods for differential gene expression (DGE) analysis (t-test and edgeR) as benchmarks, and three ML methods: Random Forests (RFs), Extreme Gradient Boosting (XGBoost), and combination of both RF and XGBoost (RX). Utility of a subset of candidate genes selected from each method for classification of FE animals was assessed by support vector machine (SVM). Among all methods, the smallest subsets of genes (117) identified by RX outperformed those chosen by t-test, edgeR, RF, or XGBoost in classification accuracy of animals. Gene co-expression network analysis confirmed the interactivity existing among these genes and their relevance within the network related to their prediction ranking based on ML. The results demonstrate a great potential for applying a combination of ML methods to large transcriptome datasets to identify biologically important genes for accurately classifying FE animals.

Mortality Prediction in Cerebral Hemorrhage Patients Using Machine Learning Algorithms in Intensive Care Units

Objectives: This study aims to investigate whether the machine learning algorithms could provide an optimal early mortality prediction method compared with other scoring systems for patients with cerebral hemorrhage in intensive care units in clinical practice.

Methods: Between 2008 and 2012, from Intensive Care III (MIMIC-III) database, all cerebral hemorrhage patients monitored with the MetaVision system and admitted to intensive care units were enrolled in this study. The calibration, discrimination, and risk classification of predicted hospital mortality based on machine learning algorithms were assessed. The primary outcome was hospital mortality. Model performance was assessed with accuracy and receiver operating characteristic curve analysis.

Results: Of 760 cerebral hemorrhage patients enrolled from MIMIC database [mean age, 68.2 years (SD, ±15.5)], 383 (50.4%) patients died in hospital, and 377 (49.6%) patients survived. The area under the receiver operating characteristic curve (AUC) of six machine learning algorithms was 0.600 (nearest neighbors), 0.617 (decision tree), 0.655 (neural net), 0.671(AdaBoost), 0.819 (random forest), and 0.725 (gcForest). The AUC was 0.423 for Acute Physiology and Chronic Health Evaluation II score. The random forest had the highest specificity and accuracy, as well as the greatest AUC, showing the best ability to predict in-hospital mortality.

Conclusions: Compared with conventional scoring system and the other five machine learning algorithms in this study, random forest algorithm had better performance in predicting in-hospital mortality for cerebral hemorrhage patients in intensive care units, and thus further research should be conducted on random forest algorithm.

Leveraging Multiple Layers of Data To Predict Drosophila Complex Traits

The ability to accurately predict complex trait phenotypes from genetic and genomic data are critical for the implementation of personalized medicine and precision agriculture; however, prediction accuracy for most complex traits is currently low. Here, we used data on whole genome sequences, deep RNA sequencing, and high quality phenotypes for three quantitative traits in the ∼200 inbred lines of the Drosophila melanogaster Genetic Reference Panel (DGRP) to compare the prediction accuracies of gene expression and genotypes for three complex traits. We found that expression levels (r = 0.28 and 0.38, for females and males, respectively) provided higher prediction accuracy than genotypes (r = 0.07 and 0.15, for females and males, respectively) for starvation resistance, similar prediction accuracy for chill coma recovery (null for both models and sexes), and lower prediction accuracy for startle response (r = 0.15 and 0.14 for female and male genotypes, respectively; and r = 0.12 and 0.11, for females and male transcripts, respectively). Models including both genotype and expression levels did not outperform the best single component model. However, accuracy increased considerably for all the three traits when we included gene ontology (GO) category as an additional layer of information for both genomic variants and transcripts. We found strongly predictive GO terms for each of the three traits, some of which had a clear plausible biological interpretation. For example, for starvation resistance in females, GO:0033500 (r = 0.39 for transcripts) and GO:0032870 (r = 0.40 for transcripts), have been implicated in carbohydrate homeostasis and cellular response to hormone stimulus (including the insulin receptor signaling pathway), respectively. In summary, this study shows that integrating different sources of information improved prediction accuracy and helped elucidate the genetic architecture of three Drosophila complex phenotypes.

Genetic mechanisms underlying feed utilization and implementation of genomic selection for improved feed efficiency in dairy cattle

The economic importance of genetically improving feed efficiency has been recognized by cattle producers worldwide. It has the potential to considerably reduce costs, minimize environmental impact, optimize land and resource use efficiency, and improve the overall cattle industry’s profitability. Feed efficiency is a genetically complex trait that can be described as units of product output (e.g., milk yield) per unit of feed input. The main objective of this review paper is to present an overview of the main genetic and physiological mechanisms underlying feed utilization in ruminants and the process towards implementation of genomic selection for feed efficiency in dairy cattle. In summary, feed efficiency can be improved via numerous metabolic pathways and biological mechanisms through genetic selection. Various studies have indicated that feed efficiency is heritable, and genomic selection can be successfully implemented in dairy cattle with a large enough training population. In this context, some organizations have worked collaboratively to do research and develop training populations for successful implementation of joint international genomic evaluations. The integration of “-omics” technologies, further investments in high-throughput phenotyping, and identification of novel indicator traits will also be paramount in maximizing the rates of genetic progress for feed efficiency in dairy cattle worldwide.

Genome-wide prediction for complex traits under the presence of dominance effects in simulated populations using GBLUP and machine learning methods

The aim of this study was to compare the predictive performance of the Genomic Best Linear Unbiased Predictor (GBLUP) and machine learning methods (Random Forest, RF; Support Vector Machine, SVM; Artificial Neural Network, ANN) in simulated populations presenting different levels of dominance effects. Simulated genome comprised 50k SNP and 300 QTL, both biallelic and randomly distributed across 29 autosomes. A total of six traits were simulated considering different values for the narrow and broad-sense heritability. In the purely additive scenario with low heritability (h2 = 0.10), the predictive ability obtained using GBLUP was slightly higher than the other methods whereas ANN provided the highest accuracies for scenarios with moderate heritability (h2 = 0.30). The accuracies of dominance deviations predictions varied from 0.180 to 0.350 in GBLUP extended for dominance effects (GBLUP-D), from 0.06 to 0.185 in RF and they were null using the ANN and SVM methods. Although RF has presented higher accuracies for total genetic effect predictions, the mean-squared error values in such a model were worse than those observed for GBLUP-D in scenarios with large additive and dominance variances. When applied to prescreen important regions, the RF approach detected QTL with high additive and/or dominance effects. Among machine learning methods, only the RF was capable to cover implicitly dominance effects without increasing the number of covariates in the model, resulting in higher accuracies for the total genetic and phenotypic values as the dominance ratio increases. Nevertheless, whether the interest is to infer directly on dominance effects, GBLUP-D could be a more suitable method.

Review: Application and Prospective Discussion of Machine Learning for the Management of Dairy Farms

Dairy farmers use herd management systems, behavioral sensors, feeding lists, breeding schedules, and health records to document herd characteristics. Consequently, large amounts of dairy data are becoming available. However, a lack of data integration makes it difficult for farmers to analyze the data on their dairy farm, which indicates that these data are currently not being used to their full potential. Hence, multiple issues in dairy farming such as low longevity, poor performance, and health issues remain. We aimed to evaluate whether machine learning (ML) methods can solve some of these existing issues in dairy farming. This review summarizes peer-reviewed ML papers published in the dairy sector between 2015 and 2020. Ultimately, 97 papers from the subdomains of management, physiology, reproduction, behavior analysis, and feeding were considered in this review. The results confirm that ML algorithms have become common tools in most areas of dairy research, particularly to predict data. Despite the quantity of research available, most tested algorithms have not performed sufficiently for a reliable implementation in practice. This may be due to poor training data. The availability of data resources from multiple farms covering longer periods would be useful to improve prediction accuracies. In conclusion, ML is a promising tool in dairy research, which could be used to develop and improve decision support for farmers. As the cow is a multifactorial system, ML algorithms could analyze integrated data sources that describe and ultimately allow managing cows according to all relevant influencing factors. However, both the integration of multiple data sources and the obtainability of public data currently remain challenging.

Predictive Features of Thymic Carcinoma and High-Risk Thymomas Using Random Forest Analysis

PURPOSE

To determine the predictive features of thymic carcinomas and high-risk thymomas using random forest algorithm.

METHODS

A total of 137 patients with pathologically confirmed high-risk thymomas and thymic carcinomas were enrolled in this study. Three clinical features and 20 computed tomography features were reviewed. The association between computed tomography features and pathological patterns was analyzed by univariate analysis and random forest. The predictive efficiency of the random forest algorithm was evaluated by receiver operating characteristic curve analysis.

RESULTS

There were 92 thymic carcinomas and 45 high-risk thymomas in this study. In univariate analysis, patient age, presence of myasthenia gravis, lesion shape, enhancement pattern, presence of necrosis or cystic change, mediastinal invasion, vessel invasion, lymphadenopathy, pericardial effusion, and distant organ metastasis were found to be statistically different between high-risk thymomas and thymic carcinomas (all P < 0.01). Random forest suggested that tumor shape, lymphadenopathy, and the presence of pericardial effusion were the key features in tumor differentiation. The predictive accuracy for the test data and whole data was 94.73% and 96.35%, respectively. Further receiver operating characteristic curve analysis showed the area under the curve was 0.957 (95% confidence interval, 0.986-0.929).

CONCLUSIONS

The random forest model in the present study has high efficiency in predictive diagnosis of thymic carcinomas and high-risk thymomas. Tumor shape, lymphadenopathy, and pericardial effusion are the key features for tumor differentiation. Thymic tumors with irregular shape, the presence of lymphadenopathy, and pericardial effusion are highly indicative of thymic carcinomas.

Identification of genetic loci associated with growth traits at weaning in yak through a genome-wide association study

A multilocus GWAS was performed to explore the genetic architecture of four growth traits in yak. In total, 354 female yaks for which measurements of body weight (BW), withers height (WH), body length (BL) and chest girth (CG) at weaning were available underwent genotyping with the Illumina BovineHD BeadChip (770K). After quality control, we retained 98 688 SNPs and 354 animals for GWAS analysis. We identified seven, 18, seven and nine SNPs (corresponding to seven, 17, seven and eight candidate genes) associated with BW, WH, BL and CG at weaning respectively. Interestingly, most of these candidate genes were reported to be involved in growth-related processes such as muscle formation, lipid deposition, feed efficiency, carcass composition and development of the central and peripheral nervous system. Our results offer novel insight into the molecular architecture underpinning yak growth traits. Further functional analyses are thus warranted to explore the molecular mechanisms whereby these genes affect these traits of interest.

Investigation of muscle transcriptomes using gradient boosting machine learning identifies molecular predictors of feed efficiency in growing pigs

Background

Improving feed efficiency (FE) is a major challenge in pig production. This complex trait is characterized by a high variability. Therefore, the identification of predictors of FE may be a relevant strategy to reduce phenotyping efforts in breeding and selection programs. The aim of this study was to investigate the suitability of expressed muscle genes in prediction of FE traits in growing pigs. The approach considered different transcriptomics experiments to cover a large range of FE values and identify reliable predictors.

Results

Microarrays data were obtained from longissimus muscles of two lines divergently selected for residual feed intake (RFI). Pigs (n = 71) from three experiments belonged to generations 6 to 8 of selection, were fed either a diet with a standard composition or a diet rich in fiber and lipids, received feed ad libitum or at restricted level, and weighed between 80 and 115 kg at slaughter. For each pig, breeding value for RFI was estimated (RFI-BV), and feed conversion ratio (FCR) and energy-based feed conversion ratio (FCRe) were calculated during the test periods. Gradient boosting algorithms were used on the merged muscle transcriptomes to identify very important predictors of FE traits. About 20,405 annotated molecular probes were commonly expressed in longissimus muscle across experiments. Six to 267 expressed muscle genes covering a variety of biological processes were found as important predictors for RFI-BV (R² = 0.63–0.65), FCR (R² = 0.61–0.70) and FCRe (R² = 0.49–0.52). The error of prediction was less than 8% for FCR. Altogether, 56 predictors were common to RFI-BV and FCR. Expression levels of 24 target genes were further measured by qPCR. Linear regression confirmed the good accuracy of combining mRNA levels of these genes to fit FE traits (RFI-BV: R² = 0.73, FRC: R² = 0.76; FCRe: R² = 0.75). Stepwise regression procedure highlighted 10 genes (FKBP5, MUM1, AKAP12, FYN, TMED3, PHKB, TGF, SOCS6, ILR4, and FRAS1) in a linear combination predicting FCR and FCRe. In addition, FKBP5 and expression levels of five other genes (IGF2, SERINC3, CSRNP3, EZR and RPL16) significantly contributed to RFI-BV.

Conclusion

It was possible to identify few genes expressed in muscle that might be reliable predictors of feed efficiency.

Electronic supplementary material

The online version of this article (10.1186/s12864-019-6010-9) contains supplementary material, which is available to authorized users.

Identification and validation of genetic variants predictive of gait in standardbred horses

Several horse breeds have been specifically selected for the ability to exhibit alternative patterns of locomotion, or gaits. A premature stop codon in the gene DMRT3 is permissive for “gaitedness” across breeds. However, this mutation is nearly fixed in both American Standardbred trotters and pacers, which perform a diagonal and lateral gait, respectively, during harness racing. This suggests that modifying alleles must influence the preferred gait at racing speeds in these populations. A genome-wide association analysis for the ability to pace was performed in 542 Standardbred horses (n = 176 pacers, n = 366 trotters) with genotype data imputed to ~74,000 single nucleotide polymorphisms (SNPs). Nineteen SNPs on nine chromosomes (ECA1, 2, 6, 9, 17, 19, 23, 25, 31) reached genome-wide significance (p < 1.44 x 10⁻⁶). Variant discovery in regions of interest was carried out via whole-genome sequencing. A set of 303 variants from 22 chromosomes with putative modifying effects on gait was genotyped in 659 Standardbreds (n = 231 pacers, n = 428 trotters) using a high-throughput assay. Random forest classification analysis resulted in an out-of-box error rate of 0.61%. A conditional inference tree algorithm containing seven SNPs predicted status as a pacer or trotter with 99.1% accuracy and subsequently performed with 99.4% accuracy in an independently sampled population of 166 Standardbreds (n = 83 pacers, n = 83 trotters). This highly accurate algorithm could be used by owners/trainers to identify Standardbred horses with the potential to race as pacers or as trotters, according to the genotype identified, prior to initiating training and would enable fine-tuning of breeding programs with designed matings. Additional work is needed to determine both the algorithm’s utility in other gaited breeds and whether any of the predictive SNPs play a physiologically functional role in the tendency to pace or tag true functional alleles.

Certain horse breeds have been developed over generations specifically for the ability to perform alternative patterns of movement, or gaits. Current understanding of the genetic basis for these gaits is limited to one known mutation apparently necessary, but not sufficient, for explaining variability in “gaitedness.” The Standardbred breed includes two distinct groups, trotters, which exhibit a two-beat gait in which the opposite forelimb and hind limb move together, and pacers, which exhibit an alternative two-beat gait where the legs on the same side of the body move together. Our long-term objective is to identify variants underlying the ability of certain Standardbreds to pace. In this study, we were able to identify several regions of the genome highly associated with pacing and, within these regions, a number of specific highly associated variants. Although the biological function of these variants has yet to be determined, we developed a model based on seven variants that was > 99% accurate in predicting whether an individual was a pacer or a trotter in two independent populations. This predictive model can be used by horse owners to make breeding and training decisions related to this economically important trait, and by scientists interested in understanding the biology of coordinated gait development.

Prediction of slaughter age in pigs and assessment of the predictive value of phenotypic and genetic information using random forest

The weight of a pig and the rate of its growth are key elements in pig production. In particular, predicting future growth is extremely useful, since it can help in determining feed costs, pen space requirements, and the age at which a pig reaches a desired slaughter weight. However, making these predictions is challenging, due to the natural variation in how individual pigs grow, and the different causes of this variation. In this paper, we used machine learning, namely random forest (RF) regression, for predicting the age at which the slaughter weight of 120 kg is reached. Additionally, we used the variable importance score from RF to quantify the importance of different types of input data for that prediction. Data of 32,979 purebred Large White pigs were provided by Topigs Norsvin, consisting of phenotypic data, estimated breeding values (EBVs), along with pedigree and pedigree-genetic relationships. Moreover, we presented a 2-step data reduction procedure, based on random projections (RPs) and principal component analysis (PCA), to extract features from the pedigree and genetic similarity matrices for use as inputs in the prediction models. Our results showed that relevant phenotypic features were the most effective in predicting the output (age at 120 kg), explaining approximately 62% of its variance (i.e., R2 = 0.62). Estimated breeding value, pedigree, or pedigree-genetic features interchangeably explain 2% of additional variance when added to the phenotypic features, while explaining, respectively, 38%, 39%, and 34% of the variance when used separately.

Development and validation of a small SNP panel for feed efficiency in beef cattle

The objective of this study was to develop and validate a customized cost-effective single nucleotide polymorphism (SNP) panel for genetic improvement of feed efficiency in beef cattle. The SNPs identified in previous association studies and through extensive analysis of candidate genomic regions and genes, were screened for their functional impact and allele frequency in Angus and Hereford breeds used as validation candidates for the panel. Association analyses were performed on genotypes of 159 SNPs from new samples of Angus (n = 160), Hereford (n = 329), and Angus-Hereford crossbred (n = 382) cattle using allele substitution and genotypic models in ASReml. Genomic heritabilities were estimated for feed efficiency traits using the full set of SNPs, SNPs associated with at least one of the traits (at P ≤ 0.05 and P < 0.10), as well as the Illumina bovine 50K representing a widely used commercial genotyping panel. A total of 63 SNPs within 43 genes showed association (P ≤ 0.05) with at least one trait. The minor alleles of SNPs located in the GHR and CAST genes were associated with decreasing effects on residual feed intake (RFI) and/or RFI adjusted for backfat (RFIf), whereas minor alleles of SNPs within MKI67 gene were associated with increasing effects on RFI and RFIf. Additionally, the minor allele of rs137400016 SNP within CNTFR was associated with increasing average daily gain (ADG). The SNPs genotypes within UMPS, SMARCAL, CCSER1, and LMCD1 genes showed significant over-dominance effects whereas other SNPs located in SMARCAL1, ANXA2, CACNA1G, and PHYHIPL genes showed additive effects on RFI and RFIf. Gene enrichment analysis indicated that gland development, as well as ion and cation transport are important physiological mechanisms contributing to variation in feed efficiency traits. The study revealed the effect of the Jak-STAT signaling pathway on feed efficiency through the CNTFR, OSMR, and GHR genes. Genomic heritability using the 63 significant (P ≤ 0.05) SNPs was 0.09, 0.09, 0.13, 0.05, 0.05, and 0.07 for ADG, dry matter intake, midpoint metabolic weight, RFI, RFIf, and backfat, respectively. These SNPs contributed to genetic variation in the studied traits and thus can potentially be used or tested to generate cost-effective molecular breeding values for feed efficiency in beef cattle.

Identification of genome regions determining semen quality in Holstein-Friesian bulls using information theory

Use of information theory can be an alternative statistical approach to detect genome regions and candidate genes that are associated with livestock traits. The aim of this study was to verify the validity of the SNPs effects on some semen quality variables of bulls using entropy analysis. Records from 288 Holstein-Friesian bulls from one AI station were included. The following semen quality variables were analyzed: CASA kinematic variables of sperm (total motility, average path velocity, straight line velocity, curvilinear velocity, amplitude of lateral head displacement, beat cross frequency, straightness, linearity), sperm membrane integrity (plazmolema, mitochondrial function), sperm ATP content. Molecular data included 48,192 SNPs. After filtering (call rate = 0.95 and MAF = 0.05), 34,794 SNPs were included in the entropy analysis. The entropy and conditional entropy were estimated for each SNP. Conditional entropy quantifies the remaining uncertainty about values of the variable with the knowledge of SNP. The most informative SNPs for each variable were determined. The computations were performed using the R statistical package. A majority of the loci had relatively small contributions. The most informative SNPs for all variables were mainly located on chromosomes: 3, 4, 5 and 16. The results from the study indicate that important genome regions and candidate genes that determine semen quality variables in bulls are located on a number of chromosomes. Some detected clusters of SNPs were located in RNA (U6 and 5S_rRNA) for all the variables for which analysis occurred. Associations between PARK2 as well GALNT13 genes and some semen characteristics were also detected.

Genome-wide association study for feed efficiency and growth traits in U.S. beef cattle

Background

Single nucleotide polymorphism (SNP) arrays for domestic cattle have catalyzed the identification of genetic markers associated with complex traits for inclusion in modern breeding and selection programs. Using actual and imputed Illumina 778K genotypes for 3887 U.S. beef cattle from 3 populations (Angus, Hereford, SimAngus), we performed genome-wide association analyses for feed efficiency and growth traits including average daily gain (ADG), dry matter intake (DMI), mid-test metabolic weight (MMWT), and residual feed intake (RFI), with marker-based heritability estimates produced for all traits and populations.

Results

Moderate and/or large-effect QTL were detected for all traits in all populations, as jointly defined by the estimated proportion of variance explained (PVE) by marker effects (PVE ≥ 1.0%) and a nominal P-value threshold (P ≤ 5e-05). Lead SNPs with PVE ≥ 2.0% were considered putative evidence of large-effect QTL (n = 52), whereas those with PVE ≥ 1.0% but < 2.0% were considered putative evidence for moderate-effect QTL (n = 35). Identical or proximal lead SNPs associated with ADG, DMI, MMWT, and RFI collectively supported the potential for either pleiotropic QTL, or independent but proximal causal mutations for multiple traits within and between the analyzed populations. Marker-based heritability estimates for all investigated traits ranged from 0.18 to 0.60 using 778K genotypes, or from 0.17 to 0.57 using 50K genotypes (reduced from Illumina 778K HD to Illumina Bovine SNP50). An investigation to determine if QTL detected by 778K analysis could also be detected using 50K genotypes produced variable results, suggesting that 50K analyses were generally insufficient for QTL detection in these populations, and that relevant breeding or selection programs should be based on higher density analyses (imputed or directly ascertained).

Conclusions

Fourteen moderate to large-effect QTL regions which ranged from being physically proximal (lead SNPs ≤ 3Mb) to fully overlapping for RFI, DMI, ADG, and MMWT were detected within and between populations, and included evidence for pleiotropy, proximal but independent causal mutations, and multi-breed QTL. Bovine positional candidate genes for these traits were functionally conserved across vertebrate species.

Electronic supplementary material

The online version of this article (doi:10.1186/s12864-017-3754-y) contains supplementary material, which is available to authorized users.

Use of genotype × environment interaction model to accommodate genetic heterogeneity for residual feed intake, dry matter intake, net energy in milk, and metabolic body weight in dairy cattle

Feed efficiency in dairy cattle has gained much attention recently. Due to the cost-prohibitive measurement of individual feed intakes, combining data from multiple countries is often necessary to ensure an adequate reference population. It may then be essential to model genetic heterogeneity when making inferences about feed efficiency or selecting efficient cattle using genomic information. In this study, we constructed a marker × environment interaction model that decomposed marker effects into main effects and interaction components that were specific to each environment. We compared environment-specific variance component estimates and prediction accuracies from the interaction model analyses, an across-environment analyses ignoring population stratification, and a within-environment analyses using an international feed efficiency data set. Phenotypes included residual feed intake, dry matter intake, net energy in milk, and metabolic body weight from 3,656 cows measured in 3 broadly defined environments: North America (NAM), the Netherlands (NLD), and Scotland (SAC). Genotypic data included 57,574 single nucleotide polymorphisms per animal. The interaction model gave the highest prediction accuracy for metabolic body weight, which had the largest estimated heritabilities ranging from 0.37 to 0.55. The within-environment model performed the best when predicting residual feed intake, which had the lowest estimated heritabilities ranging from 0.13 to 0.41. For traits (dry matter intake and net energy in milk) with intermediate estimated heritabilities (0.21 to 0.50 and 0.17 to 0.53, respectively), performance of the 3 models was comparable. Genomic correlations between environments also were computed using variance component estimates from the interaction model. Averaged across all traits, genomic correlations were highest between NAM and NLD, and lowest between NAM and SAC. In conclusion, the interaction model provided a novel way to evaluate traits measured in multiple environments in which genetic heterogeneity may exist. This model allowed estimation of environment-specific parameters and provided genomic predictions that approached or exceeded the accuracy of competing within- or across-environment models.

On the performance of tests for the detection of signatures of selection: a case study with the Spanish autochthonous beef cattle populations

Background

Procedures for the detection of signatures of selection can be classified according to the source of information they use to reject the null hypothesis of absence of selection. Three main groups of tests can be identified that are based on: (1) the analysis of the site frequency spectrum, (2) the study of the extension of the linkage disequilibrium across the length of the haplotypes that surround the polymorphism, and (3) the differentiation among populations. The aim of this study was to compare the performance of a subset of these procedures by using a dataset on seven Spanish autochthonous beef cattle populations.

Results

Analysis of the correlations between the logarithms of the statistics that were obtained by 11 tests for detecting signatures of selection at each single nucleotide polymorphism confirmed that they can be clustered into the three main groups mentioned above. A factor analysis summarized the results of the 11 tests into three canonical axes that were each associated with one of the three groups. Moreover, the signatures of selection identified with the first and second groups of tests were shared across populations, whereas those with the third group were more breed-specific. Nevertheless, an enrichment analysis identified the metabolic pathways that were associated with each group; they coincided with canonical axes and were related to immune response, muscle development, protein biosynthesis, skin and pigmentation, glucose metabolism, fat metabolism, embryogenesis and morphology, heart and uterine metabolism, regulation of the hypothalamic–pituitary–thyroid axis, hormonal, cellular cycle, cell signaling and extracellular receptors.

Conclusions

We show that the results of the procedures used to identify signals of selection differed substantially between the three groups of tests. However, they can be classified using a factor analysis. Moreover, each canonical factor that coincided with a group of tests identified different signals of selection, which could be attributed to processes of selection that occurred at different evolutionary times. Nevertheless, the metabolic pathways that were associated with each group of tests were similar, which suggests that the selection events that occurred during the evolutionary history of the populations probably affected the same group of traits.

Electronic supplementary material

The online version of this article (doi:10.1186/s12711-016-0258-1) contains supplementary material, which is available to authorized users.

Genetic and phenotypic associations of type traits and body condition score with dry matter intake, milk yield, and number of breedings in first lactation Canadian Holstein cows

The objective of the present study was to estimate genetic parameters of milk yield (MY), intake traits, type traits, body condition score (BCS), and number of breedings (NOB) in first lactation Canadian Holsteins with a focus on the possibility of using type traits as an indicator of feed intake. Data were obtained from the Canadian Dairy Network and Valacta. A mixed linear model was fitted under REML for the statistical analysis. The multivariate (five traits) model included the fixed effects of age at calving, stage of lactation, and herd-round-classifier for type traits; age at calving, stage of lactation, and herd–year–season of calving (HYS) for BCS; age at calving and HYS for MY, feed intake, and NOB. Animal and residual effects were fitted as random effects for all traits. Estimates of heritabilities for MY, dry matter intake (DMI), angularity, body depth, stature, dairy strength, final score, BCS, and NOB were 0.41, 0.13, 0.24, 0.30, 0.50, 0.30, 0.22, 0.20, and 0.02, respectively. Genetic correlations between type traits and DMI ranged from 0.16 to 0.60. Results indicate that type traits appear to have the potential to predict DMI as a combination/index of two or more traits.

Investigation of regions impacting inbreeding depression and their association with the additive genetic effect for United States and Australia Jersey dairy cattle

Background

Variation in environment, management practices, nutrition or selection objectives has led to a variety of different choices being made in the use of genetic material between countries. Differences in genome-level homozygosity between countries may give rise to regions that result in inbreeding depression to differ. The objective of this study was to characterize regions that have an impact on a runs of homozygosity (ROH) metric and estimate their association with the additive genetic effect of milk (MY), fat (FY) and protein yield (PY) and calving interval (CI) using Australia (AU) and United States (US) Jersey cows.

Methods

Genotyped cows with phenotypes on MY, FY and PY (n = 6751 US; n = 3974 AU) and CI (n = 5816 US; n = 3905 AU) were used in a two-stage analysis. A ROH statistic (ROH4Mb), which counts the frequency of a SNP being in a ROH of at least 4 Mb was calculated across the genome. In the first stage, residuals were obtained from a model that accounted for the portion explained by the estimated breeding value. In the second stage, these residuals were regressed on ROH4Mb using a single marker regression model and a gradient boosted machine (GBM) algorithm. The relationship between the additive and ROH4Mb of a region was characterized based on the (co)variance of 500 kb estimated genomic breeding values derived from a Bayesian LASSO analysis. Phenotypes to determine ROH4Mb and additive effects were residuals from the two-stage approach and yield deviations, respectively.

Results

Associations between yield traits and ROH4Mb were found for regions on BTA13, BTA23 and BTA25 for the US population and BTA3, BTA7, BTA17 for the AU population. Only one association (BTA7) was found for CI and ROH4Mb for the US population. Multiple potential epistatic interactions were characterized based on the GBM analysis. Lastly, the covariance sign between ROH4Mb and additive SNP effect of a region was heterogeneous across the genome.

Conclusion

We identified multiple genomic regions associated with ROH4Mb in US and AU Jersey females. The covariance of regions impacting ROH4Mb and the additive genetic effect were positive and negative, which provides evidence that the homozygosity effect is location dependent.

Electronic supplementary material

The online version of this article (doi:10.1186/s12864-015-2001-7) contains supplementary material, which is available to authorized users.

Exploring the areas of applicability of whole-genome prediction methods for Asian rice (Oryza sativa L.)

Our simulation results clarify the areas of applicability of nine prediction methods and suggest the factors that affect their accuracy at predicting empirical traits. Whole-genome prediction is used to predict genetic value from genome-wide markers. The choice of method is important for successful prediction. We compared nine methods using empirical data for eight phenological and morphological traits of Asian rice cultivars (Oryza sativa L.) and data simulated from real marker genotype data. The methods were genomic BLUP (GBLUP), reproducing kernel Hilbert spaces regression (RKHS), Lasso, elastic net, random forest (RForest), Bayesian lasso (Blasso), extended Bayesian lasso (EBlasso), weighted Bayesian shrinkage regression (wBSR), and the average of all methods (Ave). The objectives were to evaluate the predictive ability of these methods in a cultivar population, to characterize them by exploring the area of applicability of each method using simulation, and to investigate the causes of their different accuracies for empirical traits. GBLUP was the most accurate for one trait, RKHS and Ave for two, and RForest for three traits. In the simulation, Blasso, EBlasso, and Ave showed stable performance across the simulated scenarios, whereas the other methods, except wBSR, had specific areas of applicability; wBSR performed poorly in most scenarios. For each method, the accuracy ranking for the empirical traits was largely consistent with that in one of the simulated scenarios, suggesting that the simulation conditions reflected the factors that affected the method accuracy for the empirical results. This study will be useful for genomic prediction not only in Asian rice, but also in populations from other crops with relatively small training sets and strong linkage disequilibrium structures.

Short communication: Use of single nucleotide polymorphism genotypes and health history to predict future phenotypes for milk production, dry matter intake, body weight, and residual feed intake in dairy cattle

As feed prices have increased, the efficiency of feed utilization in dairy cattle has attracted increasing attention. In this study, we used residual feed intake (RFI) as a measurement of feed efficiency along with its component traits, adjusted milk energy (aMilkE), adjusted dry matter intake (aDMI), and adjusted metabolic body weight (aMBW), where the adjustment was for environmental factors. These traits may also be affected by prior health problems. Therefore, the carryover effects of 3 health traits from the rearing period and 10 health traits from the lactating period (in the same lactation before phenotype measurements) on RFI, aMilkE, aDMI, and aMBW were evaluated. Cows with heavier birth weight and greater body weight at calving of this lactation had significant increases in aMilkE, aDMI, and aMBW. The only trait associated with RFI was the incidence of diarrhea early in the lactation. Mastitis and reproductive problems had negative carryover effects on aMilkE. The aMBW of cows with metabolic disorders early in the lactation was lower than that of unaffected cows. The incidence of respiratory disease during lactating period was associated with greater aMBW and higher aDMI. To examine the contribution of health traits to the accuracy of predicted phenotype, genomic predictions were computed with or without information regarding 13 health trait phenotypes using random forests (RF) and support vector machine algorithms. Adding health trait phenotypes increased prediction accuracies slightly, except for prediction of RFI using RF. In general, the accuracies were greater for support vector machine than RF, especially for RFI. The methods described herein can be used to predict future phenotypes for dairy replacement heifers, thereby facilitating culling decisions that can lead to decreased feed costs during the rearing period. For these decisions, prediction of the animal's own phenotype is of greater importance than prediction of the genetic superiority or inferiority that will transmit to its offspring.

Bovine NR1I3 gene polymorphisms and its association with feed efficiency traits in Nellore cattle

The Nuclear receptor 1 family I member 3 (NR1I3), also known as the Constitutive Androstane Receptor (CAR), was initially characterized as a key regulator of xenobiotic metabolism. However, recent biochemical and structural data suggest that NR1I3 is activated in response to metabolic and nutritional stress in a ligand-independent manner. Thus, we prospected the Bovine NR1I3 gene for polymorphisms and studied their association with feed efficiency traits in Nellore cattle. First, 155 purebred Nellore bulls were individually measured for Residual Feed Intake (RFI) and the 25 best (High Feed Efficiency group, HFE) and the 25 worst animals (Low Feed Efficiency group, LFE) were selected for DNA extraction. The entire Bovine NR1I3 gene was amplified and polymorphisms were identified by sequencing. Then, one SNP different between HFE and LFE groups was genotyped in all the 155 animals and in another 288 animals totalizing 443 Nellore bulls genotyped for association of NR1I3 SNPs with feed efficiency traits. We found 24 SNPs in the NR1I3 gene and choose a statistically different SNP between HFE and LFE groups for further analysis. Genotyping of the 155 animals showed a significant association within SNP and RFI (p = 0.04), Residual Intake and BW Gain (p = 0.04) and Dry Matter Intake (p = 0.01). This SNP is located in the 5′flanking promoter region of NR1I3 gene and different alleles alter the binding site for predicted transcriptional factors as HNF4alpha, CREM and c-MYB, leading us to conclude that NR1I3 expression and regulation might be important to feed efficiency.

Genome-wide association and systems genetic analyses of residual feed intake, daily feed consumption, backfat and weight gain in pigs

Background

Feed efficiency is one of the major components determining costs of animal production. Residual feed intake (RFI) is defined as the difference between the observed and the expected feed intake given a certain production. Residual feed intake 1 (RFI1) was calculated based on regression of individual daily feed intake (DFI) on initial test weight and average daily gain. Residual feed intake 2 (RFI2) was as RFI1 except it was also regressed with respect to backfat (BF). It has been shown to be a sensitive and accurate measure for feed efficiency in livestock but knowledge of the genomic regions and mechanisms affecting RFI in pigs is lacking. The study aimed to identify genetic markers and candidate genes for RFI and its component traits as well as pathways associated with RFI in Danish Duroc boars by genome-wide associations and systems genetic analyses.

Results

Phenotypic and genotypic records (using the Illumina Porcine SNP60 BeadChip) were available on 1,272 boars. Fifteen and 12 loci were significantly associated (p < 1.52 × 10^-6) with RFI1 and RFI2, respectively. Among them, 10 SNPs were significantly associated with both RFI1 and RFI2 implying the existence of common mechanisms controlling the two RFI measures. Significant QTL regions for component traits of RFI (DFI and BF) were detected on pig chromosome (SSC) 1 (for DFI) and 2 for (BF). The SNPs within MAP3K5 and PEX7 on SSC 1, ENSSSCG00000022338 on SSC 9, and DSCAM on SSC 13 might be interesting markers for both RFI measures. Functional annotation of genes in 0.5 Mb size flanking significant SNPs indicated regulation of protein and lipid metabolic process, gap junction, inositol phosphate metabolism and insulin signaling pathway are significant biological processes and pathways for RFI, respectively.

Conclusions

The study detected novel genetic variants and QTLs on SSC 1, 8, 9, 13 and 18 for RFI and indicated significant biological processes and metabolic pathways involved in RFI. The study also detected novel QTLs for component traits of RFI. These results improve our knowledge of the genetic architecture and potential biological pathways underlying RFI; which would be useful for further investigations of key candidate genes for RFI and for development of biomarkers.

Including dominance effects in the genomic BLUP method for genomic evaluation

We evaluated the performance of GBLUP including dominance genetic effect (GBLUP-D) by estimating variances and predicting genetic merits in a computer simulation and 2 actual traits (T4 and T5) in pigs. In simulation data, GBLUP-D explained more than 50% of dominance genetic variance. Moreover, GBLUP-D yielded estimated total genetic effects over 1.2% more accurate than those yielded by GBLUP. In particular, when the dominance genetic variance was large, the accuracy could be substantially improved by increasing the number of markers. The dominance genetic variances in T4 and T5 accounted for 9.6% and 6.3% of the phenotypic variances, respectively. Estimates of such small dominance genetic variances contributed little to the improvement of the accuracies of estimated total genetic effects. In both simulation and pig data, there were nearly no differences in the estimates of additive genetic effects or their variance between GBLUP-D and GBLUP. Therefore, we conclude GBLUP-D is a feasible approach to improve genetic performance in crossbred populations with large dominance genetic variation and identify mating systems with good combining ability.