Models to estimate genetic gain of soybean seed yield from annual multi-environment field trials

Phenotypic plasticity of flowering time and plant height related traits in wheat

BACKGROUND

Climate changes pose challenges to crop production. However, the causes of phenotypic differences across environments remain unclear.

RESULTS

Here, heading date (HD), flowering date (FD), and plant height (PH) were measured along with four environmental factors (day length (DL), growing degree days (GDD), precipitation (PRCP), and photothermal ratio (PTR)) to investigate the genetic basis of phenotypic plasticity of these traits in 616 wheat accessions using genome-wide association studies. Regarding quantitative trait locus-by-environment interactions (QEIs), five known and three candidate genes for HD, six known and seven candidate genes for FD, and four known and eighteen candidate genes for PH were identified. For the genes associated with phenotypic plasticity, 10 genes exhibited responsiveness to alterations in diverse environmental conditions according to transcriptome data; haplotype effects of 33 genes were identified as significantly correlated with the changes in environmental factors; six candidate genes were identified as hub genes in the gene network, possibly influencing other genes and causing the phenotypic plasticity. And over-dominant effects can explain over 50% the genetic variance of phenotypic plasticity. More importantly, one FD/HD candidate gene (TraesCS4A01G180700) and two PH candidate genes (TraesCS5B01G054800 and TraesCS2A01G539400) partly explain the phenotypic plasticity for the FD/HD and PH traits, respectively. In addition, the potential utilization of these genes in wheat breeding was discussed.

CONCLUSIONS

This study elucidated the genetic basis of phenotypic differences caused by environments and provided a foundation for addressing the impact of climate change on crop production.

Robust soybean seed yield estimation using high-throughput ground robot videos

We present a novel method for soybean [Glycine max (L.) Merr.] yield estimation leveraging high-throughput seed counting via computer vision and deep learning techniques. Traditional methods for collecting yield data are labor-intensive, costly, and prone to equipment failures at critical data collection times and require transportation of equipment across field sites. Computer vision, the field of teaching computers to interpret visual data, allows us to extract detailed yield information directly from images. By treating it as a computer vision task, we report a more efficient alternative, employing a ground robot equipped with fisheye cameras to capture comprehensive videos of soybean plots from which images are extracted in a variety of development programs. These images are processed through the P2PNet-Yield model, a deep learning framework, where we combined a feature extraction module (the backbone of the P2PNet-Soy) and a yield regression module to estimate seed yields of soybean plots. Our results are built on 2 years of yield testing plot data-8,500 plots in 2021 and 650 plots in 2023. With these datasets, our approach incorporates several innovations to further improve the accuracy and generalizability of the seed counting and yield estimation architecture, such as the fisheye image correction and data augmentation with random sensor effects. The P2PNet-Yield model achieved a genotype ranking accuracy score of up to 83%. It demonstrates up to a 32% reduction in time to collect yield data as well as costs associated with traditional yield estimation, offering a scalable solution for breeding programs and agricultural productivity enhancement.

Genomic and phenomic prediction for soybean seed yield, protein, and oil

Developments in genomics and phenomics have provided valuable tools for use in cultivar development. Genomic prediction (GP) has been used in commercial soybean [Glycine max L. (Merr.)] breeding programs to predict grain yield and seed composition traits. Phenomic prediction (PP) is a rapidly developing field that holds the potential to be used for the selection of genotypes early in the growing season. The objectives of this study were to compare the performance of GP and PP for predicting soybean seed yield, protein, and oil. We additionally conducted genome-wide association studies (GWAS) to identify significant single-nucleotide polymorphisms (SNPs) associated with the traits of interest. The GWAS panel of 292 diverse accessions was grown in six environments in replicated trials. Spectral data were collected at two time points during the growing season. A genomic best linear unbiased prediction (GBLUP) model was trained on 269 accessions, while three separate machine learning (ML) models were trained on vegetation indices (VIs) and canopy traits. We observed that PP had a higher correlation coefficient than GP for seed yield, while GP had higher correlation coefficients for seed protein and oil contents. VIs with high feature importance were used as covariates in a new GBLUP model, and a new random forest model was trained with the inclusion of selected SNPs. These models did not outperform the original GP and PP models. These results show the capability of using ML for in-season predictions for specific traits in soybean breeding and provide insights on PP and GP inclusions in breeding programs.

A framework for simulating genotype-by-environment interaction using multiplicative models

KEY MESSAGE

The simulation of genotype-by-environment interaction using multiplicative models provides a general and scalable framework to generate realistic multi-environment datasets and model plant breeding programmes. Plant breeding has been historically shaped by genotype-by-environment interaction (GEI). Despite its importance, however, many current simulations do not adequately capture the complexity of GEI inherent to plant breeding. The framework developed in this paper simulates GEI with desirable structure using multiplicative models. The framework can be used to simulate a hypothetical target population of environments (TPE), from which many different multi-environment trial (MET) datasets can be sampled. Measures of variance explained and expected accuracy are developed to tune the simulation of non-crossover and crossover GEI and quantify the MET-TPE alignment. The framework has been implemented within the R package FieldSimR, and is demonstrated here using two working examples supported by R code. The first example embeds the framework into a linear mixed model to generate MET datasets with low, moderate and high GEI, which are used to compare several popular statistical models applied to plant breeding. The prediction accuracy generally increases as the level of GEI decreases or the number of environments sampled in the MET increases. The second example integrates the framework into a breeding programme simulation to compare genomic and phenotypic selection strategies over time. Genomic selection outperforms phenotypic selection by ∼ 50-70% in the TPE, depending on the level of GEI. These examples demonstrate how the new framework can be used to generate realistic MET datasets and model plant breeding programmes that better reflect the complexity of real-world settings, making it a valuable tool for optimising a wide range of breeding methodologies.

Understanding the Molecular Regulatory Networks of Seed Size in Soybean

Soybean being a major cash crop provides half of the vegetable oil and a quarter of the plant proteins to the global population. Seed size traits are the most important agronomic traits determining the soybean yield. These are complex traits governed by polygenes with low heritability as well as are highly influenced by the environment as well as by genotype x environment interactions. Although, extensive efforts have been made to unravel the genetic basis and molecular mechanism of seed size in soybean. But most of these efforts were majorly limited to QTL identification, and only a few genes for seed size were isolated and their molecular mechanism was elucidated. Hence, elucidating the detailed molecular regulatory networks controlling seed size in soybeans has been an important area of research in soybeans from the past decades. This paper describes the current progress of genetic architecture, molecular mechanisms, and regulatory networks for seed sizes of soybeans. Additionally, the main problems and bottlenecks/challenges soybean researchers currently face in seed size research are also discussed. This review summarizes the comprehensive and systematic information to the soybean researchers regarding the molecular understanding of seed size in soybeans and will help future research work on seed size in soybeans.