A review of machine learning models applied to genomic prediction in animal breeding

Breaking down data silos across companies to train genome-wide predictions: A feasibility study in wheat

Big data, combined with artificial intelligence (AI) techniques, holds the potential to significantly enhance the accuracy of genome-wide predictions. Motivated by the success reported for wheat hybrids, we extended the scope to inbred lines by integrating phenotypic and genotypic data from four commercial wheat breeding programs. Acting as an academic data trustee, we merged these data with historical experimental series from previous public-private partnerships. The integrated data spanned 12 years, 168 environments, and provided a genomic prediction training set of up to ~9500 genotypes for grain yield, plant height and heading date. Despite the heterogeneous phenotypic and genotypic data, we were able to obtain high-quality data by implementing rigorous data curation, including SNP imputation. We utilized the data to compare genomic best linear unbiased predictions with convolutional neural network-based genomic prediction. Our analysis revealed that we could flexibly combine experimental series for genomic prediction, with prediction ability steadily improving as the training set sizes increased, peaking at around 4000 genotypes. As training set sizes were further increased, the gains in prediction ability decreased, approaching a plateau well below the theoretical limit defined by the square root of the heritability. Potential avenues, such as designed training sets or novel non-linear prediction approaches, could overcome this plateau and help to more fully exploit the high-value big data generated by breaking down data silos across companies.

ExAutoGP: Enhancing Genomic Prediction Stability and Interpretability with Automated Machine Learning and SHAP

Machine learning has attracted much attention in the field of genomic prediction due to its powerful predictive capabilities, yet the lack of an explanatory nature in modeling decisions remains a major challenge. In this study, we propose a novel machine learning method, ExAutoGP, which aims to improve the accuracy of genomic prediction and enhance the transparency of the model by combining automated machine learning (AutoML) with SHapley Additive exPlanations (SHAP). To evaluate ExAutoGP's effectiveness, we designed a comparative experiment consisting of a simulated dataset and two real animal datasets. For each dataset, we applied ExAutoGP and five baseline models-Genomic Best Linear Unbiased Prediction (GBLUP), BayesB, Support Vector Regression (SVR), Kernel Ridge Regression (KRR), and Random Forest (RF). All models were trained and evaluated using five repeated five-fold cross-validation, and their performance was assessed based on both predictive accuracy and computational efficiency. The results show that ExAutoGP exhibits robust and excellent prediction performance on all datasets. In addition, the SHAP method not only effectively reveals the decision-making process of ExAutoGP and enhances its interpretability, but also identifies genetic markers closely related to the traits. This study demonstrates the strong potential of AutoML in genomic prediction, while the introduction of SHAP provides actionable biological insights. The synergy of high prediction accuracy and interpretability offers new perspectives for optimizing genomic selection strategies in livestock and poultry breeding.

Review: Genomic selection in the era of phenotyping based on digital images

Promoting sustainable breeding programs requires several measures, including genomic selection and continuous data recording. Digital phenotyping uses images, videos, and sensor data to continuously monitor animal activity and behaviors, such as feeding, walking, and distress, while also measuring production traits like average daily gain, loin depth, and backfat thickness. Coupled with machine learning techniques, any feature of interest can be extracted and used as phenotypes in genomic prediction models. It can also help define novel phenotypes that are hard or expensive for humans to measure. For the already recorded traits, it may add extra precision or lower phenotyping costs. One example is lameness in pigs, where digital phenotyping has allowed moving from a categorical scoring system to a continuous phenotypic scale, resulting in increased heritability and greater selection potential. Additionally, digital phenotyping offers an effective approach for generating large datasets on difficult-to-measure behavioral traits at any given time, enabling the quantification and understanding of their relationships with production traits, which may be recorded at a less frequent basis. One example is the strong, negative genetic correlation between distance traveled and average daily gain in pigs. Conversely, despite improvements in computer vision, phenotype accuracy may not be maximized for some production or carcass traits. In this review, we discuss various image processing techniques to prepare the data for the genomic evaluation models, followed by a brief description of object detection and segmentation methodology, including model selection and objective-specific modifications to the state-of-the-art models. Then, we present real-life applications of digital phenotyping for various species, and finally, we provide further challenges. Overall, digital phenotyping is a promising tool to increase the rates of genetic gain, promote sustainable genomic selection, and lower phenotyping costs. We foresee a massive inclusion of digital phenotypes into breeding programs, making it the primary phenotyping tool.

Integration of machine learning and genome-wide association study to explore the genomic prediction accuracy of agronomic trait in oats (Avena sativa L.)

Machine learning (ML) has garnered significant attention for its potential to enhance the accuracy of genomic predictions (GPs) in various economic crops with the use of complete genomic information. Genome-wide association studies (GWAS) are widely used to pinpoint trait-related causal variant loci in genomes. However, the simultaneous integration of both methods for crop genome prediction necessitates further research. In this study, we integrated ML and GWAS to assess the efficiency of GP for seven key agronomic traits in 195 oat (Avena sativa) cultivars from major oat-growing regions around the world. A total of 94 trait-associated single nucleotide polymorphisms were identified through the GWAS study. GP studies were conducted using the classical model genomic best linear unbiased prediction (GBLUP) and six ML models. GBLUP performed poorly in predicting all traits except flag leaf width, while none of the ML models consistently provided the best prediction accuracy across all traits. The prediction accuracy of the GWAS-derived markers was better than that of the use of genome-wide markers, and plant height had the highest prediction rate at 100 GWAS-derived markers, and the rest of the traits for which more markers were required. These results play an important role in advancing the use of GP in small oat breeding programs by optimizing the prediction rate of GP and reducing the number of markers, confirming that high prediction rates can be achieved with smaller datasets.

Enhancing Genomic Prediction Accuracy of Reproduction Traits in Rongchang Pigs Through Machine Learning

The increasing volume of genome sequencing data presents challenges for traditional genome-wide prediction methods in handling large datasets. Machine learning (ML) techniques, which can process high-dimensional data, offer promising solutions. This study aimed to find a genome-wide prediction method for local pig breeds, using 10 datasets with varying SNP densities derived from imputed sequencing data of 515 Rongchang pigs and the Pig QTL database. Three reproduction traits-litter weight, total number of piglets born, and number of piglets born alive-were predicted using six traditional methods and five ML methods, including kernel ridge regression, random forest, Gradient Boosting Decision Tree (GBDT), Light Gradient Boosting Machine, and Adaboost. The methods' efficacy was evaluated using fivefold cross-validation and independent tests. The predictive performance of both traditional and ML methods initially increased with SNP density, peaking at 800-900 k SNPs. ML methods outperformed traditional ones, showing improvements of 0.4-4.1%. The integration of GWAS and the Pig QTL database enhanced ML robustness. ML models exhibited superior generalizability, with high correlation coefficients (0.935-0.998) between cross-validation and independent test results. GBDT and random forest showed high computational efficiency, making them promising methods for genomic prediction in livestock breeding.

Performance Comparison of Genomic Best Linear Unbiased Prediction and Four Machine Learning Models for Estimating Genomic Breeding Values in Working Dogs

This study investigates the efficacy of various genomic prediction models-Genomic Best Linear Unbiased Prediction (GBLUP), Random Forest (RF), Support Vector Machine (SVM), Extreme Gradient Boosting (XGB), and Multilayer Perceptron (MLP)-in predicting genomic breeding values (gEBVs). The phenotypic data include three binary health traits (anodontia, distichiasis, oral papillomatosis) and one behavioral trait (distraction) in a population of guide dogs. These traits impact the potential for success in guide dogs and are therefore routinely characterized but were chosen based on differences in heritability and case counts specifically to assess gEBV model performance. Utilizing a dataset from The Seeing Eye organization, which includes German Shepherds (n = 482), Golden Retrievers (n = 239), Labrador Retrievers (n = 1188), and Labrador and Golden Retriever crosses (n = 111), we assessed model performance within and across different breeds, trait heritability, case counts, and SNP marker densities. Our results indicate that no significant differences were found in model performance across varying heritabilities, case counts, or SNP densities, with all models performing similarly. Given its lack of need for parameter optimization, GBLUP was the most efficient model. Distichiasis showed the highest overall predictive performance, likely due to its higher heritability, while anodontia and distraction exhibited moderate accuracy, and oral papillomatosis had the lowest accuracy, correlating with its low heritability. These findings underscore that lower density SNP datasets can effectively construct gEBVs, suggesting that high-cost, high-density genotyping may not always be necessary. Additionally, the similar performance of all models indicates that simpler models like GBLUP, which requires less fine tuning, may be sufficient for genomic prediction in canine breeding programs. The research highlights the importance of standardized phenotypic assessments and carefully constructed reference populations to optimize the utility of genomic selection in canine breeding programs.

The impact of dietary fiber on colorectal cancer patients based on machine learning

Objective

This study aimed to evaluate the impact of enteral nutrition with dietary fiber on patients undergoing laparoscopic colorectal cancer (CRC) surgery.

Methods

Between January 2023 and August 2024, 164 CRC patients were randomly assigned to two groups at our hospital. The control group received standard nutritional intervention, while the observation group received enteral nutritional support containing dietary fiber. Both groups were subjected to intervention and continuously observed until the 14th postoperative day. An observational analysis assessed the impact of dietary fiber intake on postoperative nutritional status in CRC patients. The study compared infection stress index, inflammatory factors, nutritional status, intestinal function recovery, and complication incidence between groups. Additionally, four machine learning models-Logistic Regression (LR), Random Forest (RF), Neural Network (NN), and Support Vector Machine (SVM)-were developed based on nutritional and clinical indicators.

Results

In the observation group, levels of procalcitonin (PCT), beta-endorphin (β-EP), C-reactive protein (CRP), interleukin-1 (IL-1), interleukin-8 (IL-8), and tumor necrosis factor-alpha (TNF-α) were significantly lower compared to the control group (p < 0.01). Conversely, levels of albumin (ALB), hemoglobin (HB), transferrin (TRF), and prealbumin (PAB) in the observation group were significantly higher than those in the control group (p < 0.01). Furthermore, LR, RF, NN, and SVM models can effectively predict the effects of dietary fiber on the immune function and inflammatory response of postoperative CRC patients, with the NN model performing the best. Through the screening of machine learning models, four key predictors for CRC patients were identified: PCT, PAB, ALB, and IL-1.

Conclusion

Postoperative dietary fiber administration in colorectal cancer enhances immune function, reduces disease-related inflammation, and inhibits tumor proliferation. Machine learning-based CRC prediction models hold clinical value.

An investigation of machine learning methods applied to genomic prediction in yellow-feathered broilers

Machine learning (ML) methods have rapidly developed in various theoretical and practical research areas, including predicting genomic breeding values for large livestock animals. However, few studies have investigated the application of ML in broiler breeding. In this study, seven different ML methods-support vector regression (SVR), random forest (RF), gradient boosting decision tree (GBDT), extreme gradient boosting (XGBoost), light gradient boosting machine (LightGBM), kernel ridge regression (KRR) and multilayer perceptron (MLP) were employed to predict the genomic breeding values of laying traits, growth and carcass traits in a yellow-feathered broiler breeding population. The results indicated that classic methods, such as GBLUP and Bayesian, achieved superior prediction accuracy compared to ML methods in five of the eight traits. For half-eviscerated weight (HEW), ML methods showed an average improvement of 54.4% over GBLUP and Bayesian methods. Among the ML methods, SVR, RF, GBDT, and XGBoost exhibited improvements exceeding 60%, with respective values of 61.3%, 61.0%, 60.4%, and 60.7%; while MLP improved by 54.4% and LightGBM by 53.7%, KRR had the lowest improvement at 29.4%. For eviscerated weight (EW), ML methods still outperformed GBLUP and Bayesian methods. MLP gained the largest improvement at 19.0%, while SVR, RF, GBDT, XGBoost, LightGBM, and KRR improved by 15.0%, 16.5%, 9.5%, 7.0%, 1.6%, and 15.9%, respectively. Compared to default hyperparameters, the average improvement of ML methods with tuned hyperparameters was 34.0%, 32.9%, 27.0%, 19.3%, 26.8%, 13.2%, 18.9%, and 46.3%, respectively. The prediction accuracy of above algorithms could be optimized using genome-wide association study (GWAS) to select subsets of significant SNPs. This work provides valuable insights into genomic prediction, aiding genetic breeding in broilers.

Sex identification in rainbow trout using genomic information and machine learning

Sex identification in farmed fish is important for the management of fish stocks and breeding programs, but identification based on visual characteristics is typically difficult or impossible in juvenile or premature fish. The amount of genomic data obtained from farmed fish is rapidly growing with the implementation of genomic selection in aquaculture. In comparison to mammals and birds, ray-finned fishes exhibit a greater diversity of sex determination systems, with an absence of conserved genomic regions. A group of genomic markers located on a standard genotyping array has been reported to potentially be linked with sex determination in rainbow trout. However, the set of markers suitable for sex identification may vary between populations. Sex identification from genomic data is usually performed using probabilistic methods, where suitable markers are known beforehand. In our study, we demonstrated the use of the Extreme Gradient Boosting approach from the supervised machine learning gradient boost framework to predict sex from unimputed genomic data, when the suitability of the markers was unknown a priori. The accuracy of the method was assessed using four simulated datasets with different genotyping error rates and one real dataset from the Finnish Rainbow Trout Breeding Program. The method showed high prediction quality on both simulated and real datasets. For simulated datasets with low (5%) and high (50%) genotyping error rates, the accuracies were 1.0 and 0.60, respectively. In the real data, the method achieved a prediction accuracy of 98%, which is suitable for routine use.

Identifying Significant SNPs of the Total Number of Piglets Born and Their Relationship with Leg Bumps in Pigs

The aim of this study was to identify genetic variants and pathways associated with the total number of piglets born and to investigate the potential negative consequences of the intensive selection for reproductive traits, particularly the formation of bumps on the legs of pigs. We used genome-wide association analysis and methods for identifying selection signatures. As a result, 47 SNPs were identified, localized in genes that play a significant role during sow pregnancy. These genes are involved in follicle growth and development (SGC), early embryonic development (CCDC3, LRRC8C, LRFN3, TNFRSF19), endometrial receptivity and implantation (NEBL), placentation, and embryonic development (ESRRG, GHRHR, TUSC3, NBAS). Several genes are associated with disorders of the nervous system and brain development (BCL11B, CDNF, ULK4, CC2D2A, KCNK2). Additionally, six SNPs are associated with the formation of bumps on the legs of pigs. These variants include intronic variants in the CCDC3, ULK4, and MINDY4 genes, as well as intergenic variants, regulatory region variants, and variants in the exons of non-coding transcripts. The results suggest important biological pathways and genetic variants associated with sow fertility and highlight the potential negative impacts on the health and physical condition of pigs.

Machine learning approaches and genetic determinants that influence the development of type 2 diabetes mellitus: a genetic association study in Brazilian patients

This genetic association study including 120 patients with type 2 diabetes mellitus (T2DM) and 166 non-diabetic individuals aimed to investigate the association of polymorphisms in the genes GSTM1 and GSTT1 (gene deletion), GSTP1 (rs1695), ACE (rs4646994), ACE2 (rs2285666), VEGF-A (rs28357093), and MTHFR (rs1801133) with the development of T2DM in the population of Goiás, Brazil. Additionally, the combined effects of these polymorphisms and the possible differences between sexes in susceptibility to the disease were evaluated. Finally, machine learning models were integrated to select the main risk characteristics for the T2DM diagnosis. Risk associations were found for the GSTT1-null genotype in the non-stratified sample and females, and for mutant C allele of the VEGF-A rs28357093 polymorphism in the non-stratified sample. Furthermore, an association of heterozygous (AG) and mutant (GG) GSTP1 genotypes was observed when combined with GSTT1-null. Machine learning approaches corroborated the results found. Therefore, these results suggested that GSTT1 and GSTP1 polymorphisms may contribute to T2DM susceptibility in a Brazilian sample.

Explainable artificial intelligence for genotype-to-phenotype prediction in plant breeding: a case study with a dataset from an almond germplasm collection

Background

Advances in DNA sequencing revolutionized plant genomics and significantly contributed to the study of genetic diversity. However, predicting phenotypes from genomic data remains a challenge, particularly in the context of plant breeding. Despite significant progress, accurately predicting phenotypes from high-dimensional genomic data remains a challenge, particularly in identifying the key genetic factors influencing these predictions. This study aims to bridge this gap by integrating explainable artificial intelligence (XAI) techniques with advanced machine learning models. This approach is intended to enhance both the predictive accuracy and interpretability of genotype-to-phenotype models, thereby improving their reliability and supporting more informed breeding decisions.

Results

This study compares several ML methods for genotype-to-phenotype prediction, using data available from an almond germplasm collection. After preprocessing and feature selection, regression models are employed to predict almond shelling fraction. Best predictions were obtained by the Random Forest method (correlation = 0.727 ± 0.020, an R 2 = 0.511 ± 0.025, and an RMSE = 7.746 ± 0.199). Notably, the application of the SHAP (SHapley Additive exPlanations) values algorithm to explain the results highlighted several genomic regions associated with the trait, including one, having the highest feature importance, located in a gene potentially involved in seed development.

Conclusions

Employing explainable artificial intelligence algorithms enhances model interpretability, identifying genetic polymorphisms associated with the shelling percentage. These findings underscore XAI's efficacy in predicting phenotypic traits from genomic data, highlighting its significance in optimizing crop production for sustainable agriculture.

An overview of recent technological developments in bovine genomics

Cattle are regarded as highly valuable animals because of their milk, beef, dung, fur, and ability to draft. The scientific community has tried a number of strategies to improve the genetic makeup of bovine germplasm. To ensure higher returns for the dairy and beef industries, researchers face their greatest challenge in improving commercially important traits. One of the biggest developments in the last few decades in the creation of instruments for cattle genetic improvement is the discovery of the genome. Breeding livestock is being revolutionized by genomic selection made possible by the availability of medium- and high-density single nucleotide polymorphism (SNP) arrays coupled with sophisticated statistical techniques. It is becoming easier to access high-dimensional genomic data in cattle. Continuously declining genotyping costs and an increase in services that use genomic data to increase return on investment have both made a significant contribution to this. The field of genomics has come a long way thanks to groundbreaking discoveries such as radiation-hybrid mapping, in situ hybridization, synteny analysis, somatic cell genetics, cytogenetic maps, molecular markers, association studies for quantitative trait loci, high-throughput SNP genotyping, whole-genome shotgun sequencing to whole-genome mapping, and genome editing. These advancements have had a significant positive impact on the field of cattle genomics. This manuscript aimed to review recent advances in genomic technologies for cattle breeding and future prospects in this field.

Genomic selection for crop improvement in fruits and vegetables: a systematic scoping review

To ensure the nutritional needs of an expanding global population, it is crucial to optimize the growing capabilities and breeding values of fruit and vegetable crops. While genomic selection, initially implemented in animal breeding, holds tremendous potential, its utilization in fruit and vegetable crops remains underexplored. In this systematic review, we reviewed 63 articles covering genomic selection and its applications across 25 different types of fruit and vegetable crops over the last decade. The traits examined were directly related to the edible parts of the crops and carried significant economic importance. Comparative analysis with WHO/FAO data identified potential economic drivers underlying the study focus of some crops and highlighted crops with potential for further genomic selection research and application. Factors affecting genomic selection accuracy in fruit and vegetable studies are discussed and suggestions made to assist in their implementation into plant breeding schemes. Genetic gain in fruits and vegetables can be improved by utilizing genomic selection to improve selection intensity, accuracy, and integration of genetic variation. However, the reduction of breeding cycle times may not be beneficial in crops with shorter life cycles such as leafy greens as compared to fruit trees. There is an urgent need to integrate genomic selection methods into ongoing breeding programs and assess the actual genomic estimated breeding values of progeny resulting from these breeding programs against the prediction models.

Supplementary Information

The online version contains supplementary material available at 10.1007/s11032-024-01497-2.

Application of Pan-Omics Technologies in Research on Important Economic Traits for Ruminants

The economic significance of ruminants in agriculture underscores the need for advanced research methodologies to enhance their traits. This review aims to elucidate the transformative role of pan-omics technologies in ruminant research, focusing on their application in uncovering the genetic mechanisms underlying complex traits such as growth, reproduction, production performance, and rumen function. Pan-omics analysis not only helps in identifying key genes and their regulatory networks associated with important economic traits but also reveals the impact of environmental factors on trait expression. By integrating genomics, epigenomics, transcriptomics, metabolomics, and microbiomics, pan-omics enables a comprehensive analysis of the interplay between genetics and environmental factors, offering a holistic understanding of trait expression. We explore specific examples of economic traits where these technologies have been pivotal, highlighting key genes and regulatory networks identified through pan-omics approaches. Additionally, we trace the historical evolution of each omics field, detailing their progression from foundational discoveries to high-throughput platforms. This review provides a critical synthesis of recent advancements, offering new insights and practical recommendations for the application of pan-omics in the ruminant industry. The broader implications for modern animal husbandry are discussed, emphasizing the potential for these technologies to drive sustainable improvements in ruminant production systems.

A review of the role of epigenetic studies for intramuscular fat deposition in beef cattle

Intramuscular fat (IMF) deposition profoundly influences meat quality and economic value in beef cattle production. Meanwhile, contemporary developments in epigenetics have opened new outlooks for understanding the molecular basics of IMF regulation, and it has become a key area of research for world scholars. Therefore, the aim of this paper was to provide insight and synthesis into the intricate relationship between epigenetic mechanisms and IMF deposition in beef cattle. The methodology involves a thorough analysis of existing literature, including pertinent books, academic journals, and online resources, to provide a comprehensive overview of the role of epigenetic studies in IMF deposition in beef cattle. This review summarizes the contemporary studies in epigenetic mechanisms in IMF regulation, high-resolution epigenomic mapping, single-cell epigenomics, multi-omics integration, epigenome editing approaches, longitudinal studies in cattle growth, environmental epigenetics, machine learning in epigenetics, ethical and regulatory considerations, and translation to industry practices from perspectives of IMF deposition in beef cattle. Moreover, this paper highlights DNA methylation, histone modifications, acetylation, phosphorylation, ubiquitylation, non-coding RNAs, DNA hydroxymethylation, epigenetic readers, writers, and erasers, chromatin immunoprecipitation followed by sequencing, whole genome bisulfite sequencing, epigenome-wide association studies, and their profound impact on the expression of crucial genes governing adipogenesis and lipid metabolism. Nutrition and stress also have significant influences on epigenetic modifications and IMF deposition. The key findings underscore the pivotal role of epigenetic studies in understanding and enhancing IMF deposition in beef cattle, with implications for precision livestock farming and ethical livestock management. In conclusion, this review highlights the crucial significance of epigenetic pathways and environmental factors in affecting IMF deposition in beef cattle, providing insightful information for improving the economics and meat quality of cattle production.