Biology subjects

Powell, O. M.

Publications and source records attributed to Powell, O. M..

9 recordsLinked to original sources

bioRxiv2026

Improved Ensemble Performance by Weight Optimisation for the Genomic Prediction of Maize Flowering Time Traits

Ensembles of multiple genomic prediction models have demonstrated improved prediction performance over the individual models contributing to the ensemble. The outperformance of ensemble models is expected from the Diversity Prediction Theorem, which states that for ensembles constructed with diverse prediction models, the ensemble prediction error becomes lower than the mean prediction error of the individual models. While a naive ensemble-average model provides baseline performance improvement by aggregating all individual prediction models with equal weights, optimising weights for each individual model could further enhance ensemble prediction performance. The weights can be optimised based on their level of informativeness regarding prediction error and diversity. Here, we evaluated weighted ensemble-average models with three possible weight optimisation approaches (linear transformation, Nelder-Mead and Bayesian) using flowering time and tillering traits from two maize nested associated mapping (NAM) datasets; TeoNAM and MaizeNAM. The three proposed weighted ensemble-average approaches improved prediction performance in several of the prediction scenarios investigated. In particular, the weighted ensemble models enhanced prediction performance when the adjusted weights differed substantially from the equal weights used by the naive ensemble models. For performance comparisons among the weighted ensembles, there was no clear superiority among the proposed approaches in both prediction accuracy and error across the prediction scenarios. Weight optimisation for ensembles warrants further investigation to explore the opportunities to improve their prediction performance; for example, integration of a weighted ensemble with a simultaneous hyperparameter tuning process may offer a promising direction for further research.

bioinformatics↗

bioRxiv2025

Ensembles of Graph Neural Networks Supervised by Genotype-to-Phenotype Structures Improved Genomic Prediction Performance

Accurate selection of favourable crop genotypes has motivated the exploration of diverse prediction algorithms for crop breeding applications. One genomic prediction method that has not been fully explored is graph attention networks (GAT). By directly analysing graphical data with the attention mechanism, GAT can incorporate the genotype-to-phenotype (G2P) structure to regularise predictions. As one potential G2P structure, a gene network can be inferred from interpretable machine learning models to effectively learn key features of prediction patterns, potentially improving prediction performance. Here, we investigated whether incorporating such data-driven prior knowledge into GAT improved prediction performance compared to GAT models representing a continuum of G2P structures, ranging from infinitesimal to fully connected. Applying the Diversity Prediction Theorem, we also combined these diverse G2P structures into an ensemble of GAT genomic prediction models to integrate complementary strengths of multiple models. The results for flowering time traits in two maize nested association mapping datasets showed a lack of consistent performance improvement in the data-driven prior knowledge GAT model. However, consistent outperformance was observed for the ensemble of GAT models. Improved predictions from the ensemble model may be driven by its ability to capture a more complete representation of the inferred gene network through the integration of information from diverse G2P structures. The observed results using the GAT methodology provided the foundation for potential performance improvement using GAT by integrating biological prior knowledge derived from omics data and empirically verified gene interactions in future research, thereby potentially enhancing the GAT ensemble performance.

genomics↗

bioRxiv2025

Ensemble-based genomic prediction for maize flowering time reveals novel insights into trait genetic architecture and improves prediction for breeding applications

While various genomic prediction models have been evaluated for their potential to accelerate genetic gain for multiple traits, no individual genomic prediction model has outperformed all others across all applications. As an alternative approach, ensembles of multiple individual genomic prediction models can be applied to utilise the complementary strengths of individual prediction models and offset the prediction errors of each. We used the EasiGP (Ensemble AnalySis with Interpretable Genomic Prediction) pipeline to investigate the performance of an ensemble approach, targeting flowering-time traits measured in two maize nested association mapping datasets. For both datasets, the ensemble-based prediction approach achieved higher prediction accuracy and lower prediction error across the flowering-time traits compared to each individual model. Multiple genomic regions known to contain key flowering-time related genes were repeatedly included as features across individual genomic prediction models, indicating the models successfully captured SNPs as features that are associated with genomic regions known to contain flowering-time genes. Although repeatability was high for some genomic regions, estimated marker effects varied across many genomic regions, suggesting that the models might also have captured different aspects of the genetic variation underlying the traits. The ensemble combination of the diverse views likely contributed to the improvement of prediction performance by the ensemble-based approach over the individual prediction models. Ensemble-based prediction can be applied to overcome limitations observed in the continuous exploration for the best individual genomic prediction models that can consistently achieve the highest prediction performance, thereby potentially contributing to improved prediction accuracy for applications in crop breeding. Article summaryThis study targets researchers interested in the performance of genomic prediction models. To demonstrate potential advantages of an ensemble of diverse individual genomic prediction models, we investigated the prediction of key flowering-time traits (days to anthesis and anthesis to silking interval) in two maize datasets. The ensemble approach consistently improved the prediction performance. The improvement was attributed to the offset of prediction errors by combining multiple different dimensions of trait genetic variation. Ensembles can lead to higher selection accuracy of desirable individuals for applications in crop breeding.

bioinformatics↗

bioRxiv2025

Circos Plots for Genome Level Interpretation of Genomic Prediction Models

Ensemble of multiple genomic prediction models have grown in popularity due to consistent prediction performance improvements in crop breeding. However, technical tools that analyse the predictive behaviour at the genome level are lacking. Here, we develop a computational tool called Ensemble AnalySis with Interpretable Genomic Prediction (EasiGP) that uses circos plots to visualise how different genomic prediction models quantify contributions of marker effects to trait phenotypes. As a demonstration of EasiGP, multiple genomic prediction models, spanning conventional statistical and machine learning algorithms, were used to infer the genetic architecture of days to anthesis (DTA) in a maize mapping population. The results indicate that genomic prediction models can capture different views of trait genetic architecture, even when their overall profiles of prediction accuracy are similar. Combinations of diverse views of the genetic architecture for the DTA trait in the TeoNAM study might explain the improved prediction performance achieved by ensembles, aligned with the implication of the Diversity Prediction Theorem. In addition to identifying well-known genomic regions contributing to the genetic architecture of DTA in maize, the ensemble of genomic prediction models highlighted several new genomic regions that have not been previously reported for DTA. Finally, different views of trait genetic architecture were observed across sub-populations, highlighting challenges for between-population genomic prediction. A deeper understanding of genomic prediction models with enhanced interpretability using EasiGP can reveal several critical findings at the genome level from the inferred genetic architecture, providing insights into the improvement of genomic prediction for crop breeding programs. Plain Language SummaryWhile an ensemble of genomic prediction models has been applied in crop breeding, the prediction mechanism has not been well-investigated due to the lack of a computational tool to interpret the predictive behaviour. It is critical to investigate prediction models at the genome level to understand how each model quantifies genomic marker effects contributing to the trait genetic architecture. Hence, we developed a computational tool, the Ensemble AnalySis with Interpretable Genomic Prediction (EasiGP), to investigate the genome features and predictive behaviours of the ensemble. Here, we demonstrate the utility of EasiGP using a maize breeding dataset. EasiGP visualised the genetic architecture from diverse interpretable genomic prediction models and identified several well-known key maize genes. EasiGP also revealed several potential new genomic regions for further investigation. EasiGP helps us investigate the trait genetic architecture that can be utilised to benefit crop breeding. Core ideasO_LIA new computational tool, EasiGP, was created to interpret multiple genomic prediction models at the genomic level C_LIO_LIEasiGP visualises the inferred trait genetic architecture from multiple genomic prediction models with circos plots C_LIO_LIAs a case study, EasiGP highlighted several well-known genes regulating the target trait, days to anthesis C_LIO_LIEasiGP can facilitate the discovery of novel genome regions underlying target traits for further investigation C_LI

bioinformatics↗

bioRxiv2024

Improvements in Prediction Performance of Ensemble Approaches for Genomic Prediction in Crop Breeding

The improvement of selection accuracy of genomic prediction is a key factor in accelerating genetic gain for crop breeding. Traditionally, efforts have focused on developing superior individual genomic prediction models. However, this approach has limitations due to the absence of a consistently "best" individual genomic prediction model, as suggested by the No Free Lunch Theorem. The No Free Lunch Theorem states that the performance of an individual prediction model is expected to be equivalent to the others when averaged across all prediction scenarios. To address this, we explored an alternative method: combining multiple genomic prediction models into an ensemble. The investigation of ensembles of prediction models is motivated by the Diversity Prediction Theorem, which indicates the prediction error of the many-model ensemble should be less than the average error of the individual models due to the diversity of predictions among the individual models. To investigate the implications of the No Free Lunch and Diversity Prediction Theorems, we developed a naive ensemble-average model, which equally weights the predicted phenotypes of individual models. We evaluated this model using two traits influencing crop yield--days to anthesis and tiller number per plant--in the Teosinte Nested Association Mapping dataset. The results show that the ensemble approach increased prediction accuracies and reduced prediction errors over individual genomic prediction models. The advantage of the ensemble was derived from the diverse predictions among the individual models, suggesting the ensemble captures a more comprehensive view of the genomic architecture of these complex traits. These results are in accordance with the expectations of the Diversity Prediction Theorem and suggest that ensemble approaches can enhance genomic prediction performance and accelerate genetic gain in crop breeding programs. Article summaryThis research targets selective breeding industries and researchers developing genomic prediction models to accelerate genetic gain in breeding programs. We applied the concept of an ensemble, combining multiple individual genomic prediction models, to predict key traits (days to anthesis and tiller number per plant) in a crop breeding dataset. Here, we show that an ensemble approach increased prediction accuracies and reduced prediction errors over individual genomic prediction models. These results indicate the potential for ensembles of multiple, diverse genomic prediction models to accelerate genetic gain in breeding programs by increasing the accuracy of selection decisions.

bioinformatics↗

bioRxiv2022

Genomic mate-allocation strategies exploiting additive and non-additive genetic effects to maximise total clonal performance in sugarcane

Mate-allocation in breeding programs can improve progeny performance by exploiting non-additive effects. Breeding decisions in the mate-allocation approach are based on predicted progeny merit rather than parental breeding value. This is particularly attractive when non-additive effects are significant, and the best-predicted progeny can be clonally propagated, for example sugarcane. We compared mate-allocation strategies that leverage non-additive and heterozygosity effects to maximise sugarcane clonal performance to schemes that use only additive effects to maximise breeding value. We used phenotypes and genotypes from a population of 2,909 clones phenotyped in Australias sugarcane breeding programs final assessment trials for three traits: tonnes of cane per hectare (TCH), commercial cane sugar (CCS), and fibre. The clones from the last generation of this data set were used as parents to simulate families from all possible crosses (1,225), each with 50 progenies. The breeding and clonal values of progeny were predicted using GBLUP (considering only additive effects) and the e-GBLUP model (incorporating additive, non-additive and heterozygosity effects). Integer linear programming was used to identify the optimal mate-allocation among the selected parents. Compared to the breeding value, the predicted progeny value of allocated crossing pairs based on clonal performance (e-GBLUP) increased by 57%, 12%, and 16% for TCH, CCS, and fibre, respectively. In our study, the mate-allocation strategy exploiting non-additive and heterozygosity effects resulted in better clonal performance. However, there was a noticeable decline in additive gain, particularly for TCH, which might have been caused by the presence of large epistatic effects. When crosses were chosen based on clonal performance for TCH, progenies inbreeding coefficients were found significantly lower than for random mating, indicating that utilising non-additive and heterozygosity effects has advantages for controlling inbreeding depression. Therefore, mate-allocation is recommended in clonal crops to improve clonal performance and reduce inbreeding.

genomics↗

bioRxiv2022

Extending the breeder's equation to take aim at the Target Population of Environments

1)A major focus for genomic prediction has been on improving trait prediction accuracy using combinations of algorithms and the training data sets available from plant breeding multi-environment trials (METs). Any improvements in prediction accuracy are viewed as pathways to improve traits in the reference population of genotypes and product performance in the target population of environments (TPE). To realise these breeding outcomes there must be a positive MET-TPE relationship that provides consistency between the trait variation expressed within the MET data sets that are used to train the genome-to-phenome (G2P) model for applications of genomic prediction and the realised trait and performance differences in the TPE for the genotypes that are the prediction targets. The strength of this MET-TPE relationship is usually assumed to be high, however it is rarely quantified. To date investigations of genomic prediction methods have not given adequate attention to quantifying the structure of the TPE and the MET-TPE relationship and its potential impact on training the G2P model for applications of genomic prediction to accelerate breeding outcomes for the on-farm TPE. We provide a perspective on the importance of the MET-TPE relationship as a key component for the design of genomic prediction methods to realize improved rates of genetic gain for the target yield, quality, stress tolerance and yield stability traits in the on-farm TPE.

genetics↗

bioRxiv2020

A Two-Part Strategy using Genomic Selection in Hybrid Crop Breeding Programs

Hybrid crop breeding programs using a two-part strategy produced the most genetic gain, but a maximum avoidance of inbreeding crossing scheme was required to increase long-term genetic gain. The two-part strategy uses outbred parents to complete multiple generations per year to reduce the generation interval of hybrid crop breeding programs. The maximum avoidance of inbreeding crossing scheme manages genetic variance by maintaining uniform contributions and inbreeding coefficients across all crosses. This study performed stochastic simulations to quantify the potential of a two-part strategy in combination with two crossing schemes to increase the rate of genetic gain in hybrid crop breeding programs. The two crossing schemes were: (i) a circular crossing scheme, and (ii) a maximum avoidance of inbreeding crossing scheme. The results from this study show that the implementation of genomic selection increased the rate of genetic gain, and that the two-part hybrid crop breeding program generated the highest genetic gain. This study also shows that the maximum avoidance of inbreeding crossing scheme increased long-term genetic gain in two-part hybrid crop breeding programs completing multiple selection cycles per year, as a result of maintaining higher levels of genetic variance over time. The flexibility of the two-part strategy offers further opportunities to integrate new technologies to further increase genetic gain in hybrid crop breeding programs, such as the use of outbred training populations. However, the practical implementation of the two-part strategy will require the development of bespoke transition strategies to fundamentally change the data, logistics, and infrastructure that underpin hybrid crop breeding programs. Key messageHybrid crop breeding programs using a two-part strategy produced the most genetic gain by using outbred parents to complete multiple generations per year. However, a maximum avoidance of inbreeding crossing scheme was required to manage genetic variance and increase long-term genetic gain.

genetics↗

bioRxiv2019

Genomic data enables genetic evaluation using data recorded on LMIC smallholder dairy farms

BackgroundGenetic evaluation is a central component of a breeding program. In advanced economies, most genetic evaluations depend on large quantities of data that are recorded on commercial farms. Large herd sizes and widespread use of artificial insemination create strong genetic connectedness that enables the genetic and environmental effects of an individual animals phenotype to be accurately separated. In contrast to this, herds are neither large nor have strong genetic connectedness in smallholder dairy production systems of many low to middle-income countries (LMIC). This limits genetic evaluation, and furthermore, the pedigree information needed for traditional genetic evaluation is typically unavailable. Genomic information keeps track of shared haplotypes rather than shared relatives. This information could capture and strengthen genetic connectedness between herds and through this may enable genetic evaluations for LMIC smallholder dairy farms. The objective of this study was to use simulation to quantify the power of genomic information to enable genetic evaluation under such conditions.\n\nResultsThe results from this study show: (i) the genetic evaluation of phenotyped cows using genomic information had higher accuracy compared to pedigree information across all breeding designs; (ii) the genetic evaluation of phenotyped cows with genomic information and modelling herd as a random effect had higher or equal accuracy compared to modelling herd as a fixed effect; (iii) the genetic evaluation of phenotyped cows from breeding designs with strong genetic connectedness had higher accuracy compared to breeding designs with weaker genetic connectedness; (iv) genomic prediction of young bulls was possible using marker estimates from the genetic evaluations of their phenotyped dams. For example, the accuracy of genomic prediction of young bulls from an average herd size of 1 (=1.58) was 0.40 under a breeding design with 1,000 sires mated per generation and a training set of 8,000 phenotyped and genotyped cows.\n\nConclusionsThis study demonstrates the potential of genomic information to be an enabling technology in LMIC smallholder dairy production systems by facilitating genetic evaluations with in-situ records collected from farms with herd sizes of four cows or less. Across a range of breeding designs, genomic data enabled accurate genetic evaluation of phenotyped cows and genomic prediction of young bulls using data sets that contained small herds with weak genetic connections. The use of smallholder dairy data in genetic evaluations would enable the establishment of breeding programs to improve in-situ germplasm and, if required, would enable the importation of the most suitable external germplasm. This could be individually tailored for each target environment. Together this would increase the productivity, profitability and sustainability of LMIC smallholder dairy production systems. However, data collection, including genomic data, is expensive and business models will need to be carefully constructed so that the costs are sustainably offset.

genetics↗