Early prediction of biomass in hybrid rye based on hyperspectral data surpasses genomic predictability in less-related breeding material

Galán, Rodrigo José; Bernal-Vasquez, Angela-Maria; Jebsen, Christian; Piepho, Hans‐Peter; Thorwarth, Patrick; Steffan, Philipp; Gordillo, Andrés; Miedaner, Thomas

doi:10.1007/s00122-021-03779-1

Cited by 24 publications

(26 citation statements)

References 71 publications

(78 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…2a). This corroborates the results of previous studies that illustrated the potential of phenomic prediction using spectral data to predict complex traits in different crops (Rincent et al 2018;Krause et al 2019;Galán et al 2021). Interestingly, we observed comparably low predictive abilities for the two disease resistance traits, particularly for yellow rust in one of the DH populations.…”

Section: Phenomic Prediction Using Nirs Data As Predictors Is Promising For Complex Traitssupporting

confidence: 92%

“…Thus, the use of spectral data as predictors could drastically increase the efficiency of selection at greatly reduced costs. Phenomic prediction based on NIRS or field-based hyperspectral data was reported for different crops and traits and shown to achieve promising predictive abilities, for example, in soybean (Parmley et al 2019 ), maize (Lane et al 2020 ), wheat (Rincent et al 2018 ; Krause et al 2019 ), rye (Galán et al 2020 , 2021 ), and sugarcane (Gonçalves et al 2021 ).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

The performance of phenomic selection depends on the genetic architecture of the target trait

et al. 2021

View full text Add to dashboard Cite

Key message The phenomic predictive ability depends on the genetic architecture of the target trait, being high for complex traits and low for traits with major QTL. Abstract Genomic selection is a powerful tool to assist breeding of complex traits, but a limitation is the costs required for genotyping. Recently, phenomic selection has been suggested, which uses spectral data instead of molecular markers as predictors. It was shown to be competitive with genomic prediction, as it achieved predictive abilities as high or even higher than its genomic counterpart. The objective of this study was to evaluate the performance of phenomic prediction for triticale and the dependency of the predictive ability on the genetic architecture of the target trait. We found that for traits with a complex genetic architecture, like grain yield, phenomic prediction with NIRS data as predictors achieved high predictive abilities and performed better than genomic prediction. By contrast, for mono- or oligogenic traits, for example, yellow rust, marker-based approaches achieved high predictive abilities, while those of phenomic prediction were very low. Compared with molecular markers, the predictive ability obtained using NIRS data was more robust to varying degrees of genetic relatedness between the training and prediction set. Moreover, for grain yield, smaller training sets were required to achieve a similar predictive ability for phenomic prediction than for genomic prediction. In addition, our results illustrate the potential of using field-based spectral data for phenomic prediction. Overall, our result confirmed phenomic prediction as an efficient approach to improve the selection gain for complex traits in plant breeding.

show abstract

Section: Phenomic Prediction Using Nirs Data As Predictors Is Promising For Complex Traitssupporting

confidence: 92%

Section: Introductionmentioning

confidence: 99%

The performance of phenomic selection depends on the genetic architecture of the target trait

et al. 2021

View full text Add to dashboard Cite

show abstract

“…Hyperspectral cameras are known for their multiple wavelength bands, rich spectral information, high spectral resolution, and high recognition capability. Consequently, these cameras can acquire near-continuous spectral information of features for fast, nondestructive, and high-throughput detection of crops with large data volumes [16]. Therefore, hyperspectral images are widely used in precision agriculture, such as in the monitoring for pests and diseases, estimating crop biomass, and the monitoring of crop growth [17][18][19].…”

Section: Introductionmentioning

confidence: 99%

UAV-Based Hyperspectral and Ensemble Machine Learning for Predicting Yield in Winter Wheat

et al. 2022

View full text Add to dashboard Cite

Winter wheat is a widely-grown cereal crop worldwide. Using growth-stage information to estimate winter wheat yields in a timely manner is essential for accurate crop management and rapid decision-making in sustainable agriculture, and to increase productivity while reducing environmental impact. UAV remote sensing is widely used in precision agriculture due to its flexibility and increased spatial and spectral resolution. Hyperspectral data are used to model crop traits because of their ability to provide continuous rich spectral information and higher spectral fidelity. In this study, hyperspectral image data of the winter wheat crop canopy at the flowering and grain-filling stages was acquired by a low-altitude unmanned aerial vehicle (UAV), and machine learning was used to predict winter wheat yields. Specifically, a large number of spectral indices were extracted from the spectral data, and three feature selection methods, recursive feature elimination (RFE), Boruta feature selection, and the Pearson correlation coefficient (PCC), were used to filter high spectral indices in order to reduce the dimensionality of the data. Four major basic learner models, (1) support vector machine (SVM), (2) Gaussian process (GP), (3) linear ridge regression (LRR), and (4) random forest (RF), were also constructed, and an ensemble machine learning model was developed by combining the four base learner models. The results showed that the SVM yield prediction model, constructed on the basis of the preferred features, performed the best among the base learner models, with an R2 between 0.62 and 0.73. The accuracy of the proposed ensemble learner model was higher than that of each base learner model; moreover, the R2 (0.78) for the yield prediction model based on Boruta’s preferred characteristics was the highest at the grain-filling stage.

show abstract

“…This implies that the genotypic part of NIRS estimated by multi-year analysis could be more related to the genetic signal. Thus, genetic signal ignoring genotype-by-environment interactions could be better captured when several years are combined, this was also the case in Galán et al (2021) for which multi-year spectra resulted in higher PA values.…”

Section: Nirs Variance Components and Co-inertia With Snpsmentioning

confidence: 90%

Interest of phenomic prediction as an alternative to genomic prediction in grapevine

Brault

Lazerges

Doligez

et al. 2021

Preprint

View full text Add to dashboard Cite

Phenomic prediction has been defined as an alternative to genomic prediction by using spectra instead of molecular markers. A reflectance spectrum reflects the biochemical composition within a tissue, under genetic determinism. Thus, a relationship matrix built from spectra could potentially capture genetic signal. This new methodology has been successfully applied in several cereal species but little is known so far about its interest in perennial species. Besides, phenomic prediction has only been tested for a restricted set of traits, mainly related to yield or phenology. This study aims at applying phenomic prediction for the first time in grapevine, using spectra collected on two tissues and over two consecutive years, on two populations and for 15 traits. First, we characterized the genetic signal in spectra and under which condition it could be maximized, then phenomic predictive ability was compared to genomic predictive ability. We found that the co-inertia between spectra and genomic data was stable across tissues or years, but variable across populations, with co-inertia around 0.3 and 0.6 for diversity panel and half-diallel populations, respectively. Differences between populations were also observed for predictive ability of phenomic prediction, with an average of 0.27 for the diversity panel and 0.35 for the half-diallel. For both populations, there was a correlation across traits between predictive ability of genomic and phenomic prediction, with a slope around 1 and an intercept of −0.2, thus suggesting that phenomic prediction could be applied for any trait.

show abstract

Early prediction of biomass in hybrid rye based on hyperspectral data surpasses genomic predictability in less-related breeding material

Cited by 24 publications

References 71 publications

The performance of phenomic selection depends on the genetic architecture of the target trait

The performance of phenomic selection depends on the genetic architecture of the target trait

UAV-Based Hyperspectral and Ensemble Machine Learning for Predicting Yield in Winter Wheat

Interest of phenomic prediction as an alternative to genomic prediction in grapevine

Contact Info

Product

Resources

About