Technical nuances of machine learning: implementation and validation of supervised methods for genomic prediction in plant breeding

Xavier, Alencar

doi:10.1590/1984-70332021v21sa15

Cited by 9 publications

(10 citation statements)

References 68 publications

(62 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In addition, THGS becomes an exact method that yields unbiased estimates of genetic correlations and GEBV (section Exact THGS), and reduces the bias of estimates of heritabilities, as can be demonstrated for scenario 1 (Table 6). Matrix decomposition is also useful to analyze high-dimensional datasets with many factors ( P >> N problem), and to fit one or multiple kernels of different types within multivariate ridge regression models, for example, for modeling dominance, epistasis [37], and Gaussian or Arc-cosine relationships [21,38]. The computing costs for matrix decomposition to obtain those eigenvectors, however, may outweigh the benefits for THGS as the number of individuals and markers in the analysis increases.…”

Section: Discussionmentioning

confidence: 99%

“…For balanced experimental designs, when the intercept is the only fixed effect, and either a principal components [18] or eigenvector regression [19][20][21] is used, THGS is exact. This is demonstrated in Appendix 3.…”

Section: Exact Thgsmentioning

confidence: 99%

“…Phenotypic data for five scenarios were simulated to evaluate bias and accuracy of GEBV within environments, runtime, and biases and standard errors of estimates of heritabilities and genetic correlations (Table 2). The genotypes used in the simulations came from a wheat [28][29][30][31] and a soybean dataset [21,32,33], which have been used in multiple genomic prediction studies, and are available through the R packages BGLR and SoyNAM, respectively.…”

Section: Data and Evaluation Statisticsmentioning

confidence: 99%

See 2 more Smart Citations

A new approach fits multivariate genomic prediction models efficiently

Xavier

Habier

2022

Genet Sel Evol

Self Cite

View full text Add to dashboard Cite

Background Fast, memory-efficient, and reliable algorithms for estimating genomic estimated breeding values (GEBV) for multiple traits and environments are needed to make timely decisions in breeding. Multivariate genomic prediction exploits genetic correlations between traits and environments to increase accuracy of GEBV compared to univariate methods. These genetic correlations are estimated simultaneously with GEBV, because they are specific to year, environment, and management. However, estimating genetic parameters is computationally demanding with restricted maximum likelihood (REML) and Bayesian samplers, and canonical transformations or orthogonalizations cannot be used for unbalanced experimental designs. Methods We propose a multivariate randomized Gauss–Seidel algorithm for simultaneous estimation of model effects and genetic parameters. Two previously proposed methods for estimating genetic parameters were combined with a Gauss–Seidel (GS) solver, and were called Tilde-Hat-GS (THGS) and Pseudo-Expectation-GS (PEGS). Balanced and unbalanced experimental designs were simulated to compare runtime, bias and accuracy of GEBV, and bias and standard errors of estimates of heritabilities and genetic correlations of THGS, PEGS, and REML. Models with 10 to 400 response variables, 1279 to 42,034 genetic markers, and 5990 to 1.85 million observations were fitted. Results Runtime of PEGS and THGS was a fraction of REML. Accuracies of GEBV were slightly lower than those from REML, but higher than those from the univariate approach, hence THGS and PEGS exploited genetic correlations. For 500 to 600 observations per response variable, biases of estimates of genetic parameters of THGS and PEGS were small, but standard errors of estimates of genetic correlations were higher than for REML. Bias and standard errors decreased as sample size increased. For balanced designs, GEBV and estimates of genetic correlations from THGS were unbiased when only an intercept and eigenvectors of genotype scores were fitted. Conclusions THGS and PEGS are fast and memory-efficient algorithms for multivariate genomic prediction for balanced and unbalanced experimental designs. They are scalable for increasing numbers of environments and genetic markers. Accuracy of GEBV was comparable to REML. Estimates of genetic parameters had little bias, but their standard errors were larger than for REML. More studies are needed to evaluate the proposed methods for datasets that contain selection.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Exact Thgsmentioning

confidence: 99%

Section: Data and Evaluation Statisticsmentioning

confidence: 99%

See 1 more Smart Citation

A new approach fits multivariate genomic prediction models efficiently

Xavier

Habier

2022

Genet Sel Evol

Self Cite

View full text Add to dashboard Cite

show abstract

“…Whereas decision trees are poor predictors, the collective of multiple small trees generated at random provides robust predictions. Random forest can be described by the model ( Xavier, 2021 )…”

Section: Approach 2: Multiple-marker Associationmentioning

confidence: 99%

Two decades of association mapping: Insights on disease resistance in major crops

et al. 2022

Self Cite

View full text Add to dashboard Cite

Climate change across the globe has an impact on the occurrence, prevalence, and severity of plant diseases. About 30% of yield losses in major crops are due to plant diseases; emerging diseases are likely to worsen the sustainable production in the coming years. Plant diseases have led to increased hunger and mass migration of human populations in the past, thus a serious threat to global food security. Equipping the modern varieties/hybrids with enhanced genetic resistance is the most economic, sustainable and environmentally friendly solution. Plant geneticists have done tremendous work in identifying stable resistance in primary genepools and many times other than primary genepools to breed resistant varieties in different major crops. Over the last two decades, the availability of crop and pathogen genomes due to advances in next generation sequencing technologies improved our understanding of trait genetics using different approaches. Genome-wide association studies have been effectively used to identify candidate genes and map loci associated with different diseases in crop plants. In this review, we highlight successful examples for the discovery of resistance genes to many important diseases. In addition, major developments in association studies, statistical models and bioinformatic tools that improve the power, resolution and the efficiency of identifying marker-trait associations. Overall this review provides comprehensive insights into the two decades of advances in GWAS studies and discusses the challenges and opportunities this research area provides for breeding resistant varieties.

show abstract

“…Another approach to overcome the dimensionality problem involves the use of kernels. Kernel methods have been used in plant and animal breeding for pedigree-based selection and genomic prediction (GP) (Xavier, 2021). GBLUP, the most popular kernel-based method, uses RR ('KRR') based on a linear kernel (VanRaden, 2008).…”

Section: Introductionmentioning

confidence: 99%

Genomic prediction of seasonal forage yield in perennial ryegrass

Konkolewska,

Phang,

Conaghan

et al. 2023

Grassland Research

View full text Add to dashboard Cite

BackgroundGenomic selection has the potential to accelerate genetic gain in perennial ryegrass breeding, provided complex traits such as forage yield can be predicted with sufficient accuracy.MethodsIn this study, we compared modelling approaches and feature selection strategies to evaluate the accuracy of genomic prediction models for seasonal forage yield production.ResultsOverall, model selection had limited impact on predictive ability when using the full data set. For a baseline genomic best linear unbiased prediction model, the highest mean predictive accuracy was obtained for spring grazing (0.78), summer grazing (0.62) and second cut silage (0.56). In terms of feature selection strategies, using uncorrelated single‐nucleotide polymorphisms (SNPs) had no impact on predictive ability, allowing for a potential decrease of the data set dimensions. With a genome‐wide association study, we found a significant SNP marker for spring grazing, located in the genic region annotated as coding for an enzyme responsible for fucosylation of xyloglucans—major components of the plant cell wall. We also presented an approach to increase interpretability of genomic prediction models with the use of Gene Ontology enrichment analysis.ConclusionsApproaches for feature selection will be relevant in development of low‐cost genotyping platforms in support of routine and cost‐effective implementation of genomic selection.

show abstract

Technical nuances of machine learning: implementation and validation of supervised methods for genomic prediction in plant breeding

Cited by 9 publications

References 68 publications

A new approach fits multivariate genomic prediction models efficiently

A new approach fits multivariate genomic prediction models efficiently

Two decades of association mapping: Insights on disease resistance in major crops

Genomic prediction of seasonal forage yield in perennial ryegrass

Contact Info

Product

Resources

About