Predicting grain yield using canopy hyperspectral reflectance in wheat breeding data

Osval A. Montesinos-López, Abelardo Montesinos-López, José Crossa, Gustavo los Campos, Gregorio Alvarado, Mondal Suchismita, Jessica Rutkoski, Lorena González-Pérez, Juan Burgueño

Research output: Contribution to journalArticle

Abstract

Background: Modern agriculture uses hyperspectral cameras to obtain hundreds of reflectance data measured at discrete narrow bands to cover the whole visible light spectrum and part of the infrared and ultraviolet light spectra, depending on the camera. This information is used to construct vegetation indices (VI) (e.g., green normalized difference vegetation index or GNDVI, simple ratio or SRa, etc.) which are used for the prediction of primary traits (e.g., biomass). However, these indices only use some bands and are cultivar-specific; therefore they lose considerable information and are not robust for all cultivars. Results: This study proposes models that use all available bands as predictors to increase prediction accuracy; we compared these approaches with eight conventional vegetation indexes (VIs) constructed using only some bands. The data set we used comes from CIMMYT's global wheat program and comprises 1170 genotypes evaluated for grain yield (ton/ha) in five environments (Drought, Irrigated, EarlyHeat, Melgas and Reduced Irrigated); the reflectance data were measured in 250 discrete narrow bands ranging between 392 and 851 nm. The proposed models for the simultaneous analysis of all the bands were ordinal least square (OLS), Bayes B, principal components with Bayes B, functional B-spline, functional Fourier and functional partial least square. The results of these models were compared with the OLS performed using as predictors each of the eight VIs individually and combined. Conclusions: We found that using all bands simultaneously increased prediction accuracy more than using VI alone. The Splines and Fourier models had the best prediction accuracy for each of the nine time-points under study. Combining image data collected at different time-points led to a small increase in prediction accuracy relative to models that use data from a single time-point. Also, using bands with heritabilities larger than 0.5 only in Drought as predictor variables showed improvements in prediction accuracy.

Original languageEnglish (US)
Article number4
JournalPlant Methods
Volume13
Issue number1
DOIs
StatePublished - Jan 3 2017
Externally publishedYes

Fingerprint

Least-Squares Analysis
Triticum
reflectance
Breeding
grain yield
Droughts
canopy
wheat
prediction
breeding
least squares
Ultraviolet Rays
Agriculture
cameras
Biomass
Genotype
drought
Light
cultivars
ultraviolet radiation

Keywords

  • Bayes B
  • Fourier regression
  • Genome selection
  • Prediction accuracy
  • Spectral data
  • Spline regression
  • Vegetation indexes
  • Wheat

ASJC Scopus subject areas

  • Biotechnology
  • Genetics
  • Plant Science

Cite this

Montesinos-López, O. A., Montesinos-López, A., Crossa, J., los Campos, G., Alvarado, G., Suchismita, M., ... Burgueño, J. (2017). Predicting grain yield using canopy hyperspectral reflectance in wheat breeding data. Plant Methods, 13(1), [4]. https://doi.org/10.1186/s13007-016-0154-2

Predicting grain yield using canopy hyperspectral reflectance in wheat breeding data. / Montesinos-López, Osval A.; Montesinos-López, Abelardo; Crossa, José; los Campos, Gustavo; Alvarado, Gregorio; Suchismita, Mondal; Rutkoski, Jessica; González-Pérez, Lorena; Burgueño, Juan.

In: Plant Methods, Vol. 13, No. 1, 4, 03.01.2017.

Research output: Contribution to journalArticle

Montesinos-López, OA, Montesinos-López, A, Crossa, J, los Campos, G, Alvarado, G, Suchismita, M, Rutkoski, J, González-Pérez, L & Burgueño, J 2017, 'Predicting grain yield using canopy hyperspectral reflectance in wheat breeding data', Plant Methods, vol. 13, no. 1, 4. https://doi.org/10.1186/s13007-016-0154-2
Montesinos-López OA, Montesinos-López A, Crossa J, los Campos G, Alvarado G, Suchismita M et al. Predicting grain yield using canopy hyperspectral reflectance in wheat breeding data. Plant Methods. 2017 Jan 3;13(1). 4. https://doi.org/10.1186/s13007-016-0154-2
Montesinos-López, Osval A. ; Montesinos-López, Abelardo ; Crossa, José ; los Campos, Gustavo ; Alvarado, Gregorio ; Suchismita, Mondal ; Rutkoski, Jessica ; González-Pérez, Lorena ; Burgueño, Juan. / Predicting grain yield using canopy hyperspectral reflectance in wheat breeding data. In: Plant Methods. 2017 ; Vol. 13, No. 1.
@article{f48627fbd7a041a5a886bb20925af360,
title = "Predicting grain yield using canopy hyperspectral reflectance in wheat breeding data",
abstract = "Background: Modern agriculture uses hyperspectral cameras to obtain hundreds of reflectance data measured at discrete narrow bands to cover the whole visible light spectrum and part of the infrared and ultraviolet light spectra, depending on the camera. This information is used to construct vegetation indices (VI) (e.g., green normalized difference vegetation index or GNDVI, simple ratio or SRa, etc.) which are used for the prediction of primary traits (e.g., biomass). However, these indices only use some bands and are cultivar-specific; therefore they lose considerable information and are not robust for all cultivars. Results: This study proposes models that use all available bands as predictors to increase prediction accuracy; we compared these approaches with eight conventional vegetation indexes (VIs) constructed using only some bands. The data set we used comes from CIMMYT's global wheat program and comprises 1170 genotypes evaluated for grain yield (ton/ha) in five environments (Drought, Irrigated, EarlyHeat, Melgas and Reduced Irrigated); the reflectance data were measured in 250 discrete narrow bands ranging between 392 and 851 nm. The proposed models for the simultaneous analysis of all the bands were ordinal least square (OLS), Bayes B, principal components with Bayes B, functional B-spline, functional Fourier and functional partial least square. The results of these models were compared with the OLS performed using as predictors each of the eight VIs individually and combined. Conclusions: We found that using all bands simultaneously increased prediction accuracy more than using VI alone. The Splines and Fourier models had the best prediction accuracy for each of the nine time-points under study. Combining image data collected at different time-points led to a small increase in prediction accuracy relative to models that use data from a single time-point. Also, using bands with heritabilities larger than 0.5 only in Drought as predictor variables showed improvements in prediction accuracy.",
keywords = "Bayes B, Fourier regression, Genome selection, Prediction accuracy, Spectral data, Spline regression, Vegetation indexes, Wheat",
author = "Montesinos-L{\'o}pez, {Osval A.} and Abelardo Montesinos-L{\'o}pez and Jos{\'e} Crossa and {los Campos}, Gustavo and Gregorio Alvarado and Mondal Suchismita and Jessica Rutkoski and Lorena Gonz{\'a}lez-P{\'e}rez and Juan Burgue{\~n}o",
year = "2017",
month = "1",
day = "3",
doi = "10.1186/s13007-016-0154-2",
language = "English (US)",
volume = "13",
journal = "Plant Methods",
issn = "1746-4811",
publisher = "BioMed Central",
number = "1",

}

TY - JOUR

T1 - Predicting grain yield using canopy hyperspectral reflectance in wheat breeding data

AU - Montesinos-López, Osval A.

AU - Montesinos-López, Abelardo

AU - Crossa, José

AU - los Campos, Gustavo

AU - Alvarado, Gregorio

AU - Suchismita, Mondal

AU - Rutkoski, Jessica

AU - González-Pérez, Lorena

AU - Burgueño, Juan

PY - 2017/1/3

Y1 - 2017/1/3

N2 - Background: Modern agriculture uses hyperspectral cameras to obtain hundreds of reflectance data measured at discrete narrow bands to cover the whole visible light spectrum and part of the infrared and ultraviolet light spectra, depending on the camera. This information is used to construct vegetation indices (VI) (e.g., green normalized difference vegetation index or GNDVI, simple ratio or SRa, etc.) which are used for the prediction of primary traits (e.g., biomass). However, these indices only use some bands and are cultivar-specific; therefore they lose considerable information and are not robust for all cultivars. Results: This study proposes models that use all available bands as predictors to increase prediction accuracy; we compared these approaches with eight conventional vegetation indexes (VIs) constructed using only some bands. The data set we used comes from CIMMYT's global wheat program and comprises 1170 genotypes evaluated for grain yield (ton/ha) in five environments (Drought, Irrigated, EarlyHeat, Melgas and Reduced Irrigated); the reflectance data were measured in 250 discrete narrow bands ranging between 392 and 851 nm. The proposed models for the simultaneous analysis of all the bands were ordinal least square (OLS), Bayes B, principal components with Bayes B, functional B-spline, functional Fourier and functional partial least square. The results of these models were compared with the OLS performed using as predictors each of the eight VIs individually and combined. Conclusions: We found that using all bands simultaneously increased prediction accuracy more than using VI alone. The Splines and Fourier models had the best prediction accuracy for each of the nine time-points under study. Combining image data collected at different time-points led to a small increase in prediction accuracy relative to models that use data from a single time-point. Also, using bands with heritabilities larger than 0.5 only in Drought as predictor variables showed improvements in prediction accuracy.

AB - Background: Modern agriculture uses hyperspectral cameras to obtain hundreds of reflectance data measured at discrete narrow bands to cover the whole visible light spectrum and part of the infrared and ultraviolet light spectra, depending on the camera. This information is used to construct vegetation indices (VI) (e.g., green normalized difference vegetation index or GNDVI, simple ratio or SRa, etc.) which are used for the prediction of primary traits (e.g., biomass). However, these indices only use some bands and are cultivar-specific; therefore they lose considerable information and are not robust for all cultivars. Results: This study proposes models that use all available bands as predictors to increase prediction accuracy; we compared these approaches with eight conventional vegetation indexes (VIs) constructed using only some bands. The data set we used comes from CIMMYT's global wheat program and comprises 1170 genotypes evaluated for grain yield (ton/ha) in five environments (Drought, Irrigated, EarlyHeat, Melgas and Reduced Irrigated); the reflectance data were measured in 250 discrete narrow bands ranging between 392 and 851 nm. The proposed models for the simultaneous analysis of all the bands were ordinal least square (OLS), Bayes B, principal components with Bayes B, functional B-spline, functional Fourier and functional partial least square. The results of these models were compared with the OLS performed using as predictors each of the eight VIs individually and combined. Conclusions: We found that using all bands simultaneously increased prediction accuracy more than using VI alone. The Splines and Fourier models had the best prediction accuracy for each of the nine time-points under study. Combining image data collected at different time-points led to a small increase in prediction accuracy relative to models that use data from a single time-point. Also, using bands with heritabilities larger than 0.5 only in Drought as predictor variables showed improvements in prediction accuracy.

KW - Bayes B

KW - Fourier regression

KW - Genome selection

KW - Prediction accuracy

KW - Spectral data

KW - Spline regression

KW - Vegetation indexes

KW - Wheat

UR - http://www.scopus.com/inward/record.url?scp=85010210779&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85010210779&partnerID=8YFLogxK

U2 - 10.1186/s13007-016-0154-2

DO - 10.1186/s13007-016-0154-2

M3 - Article

AN - SCOPUS:85010210779

VL - 13

JO - Plant Methods

JF - Plant Methods

SN - 1746-4811

IS - 1

M1 - 4

ER -