A simple, cost-effective high-throughput image analysis pipeline improves genomic prediction accuracy for days to maturity in wheat

Abstract

Background High-throughput phenotyping and genomic selection accelerate genetic gain in breeding programs by advances in phenotyping and genotyping methods. This study developed a simple, cost-effective high-throughput image analysis pipeline to quantify digital images taken in a panel of 286 Iran bread wheat accessions under terminal drought stress and well-watered conditions. The color proportion of green to yellow (tolerance ratio) and the color proportion of yellow to green (stress ratio) was assessed for each canopy using the pipeline. The estimated tolerance and stress ratios were used as covariates in the genomic prediction models to evaluate the effect of change in canopy color on the improvement of the genomic prediction accuracy of different agronomic traits in wheat. Results The reliability of the high-throughput image analysis pipeline was proved by three to four times of improvement in the accuracy of genomic predictions for days to maturity with the use of tolerance and stress ratios as covariates in the univariate genomic selection models. The higher prediction accuracies were attained for days to maturity when both tolerance and stress ratios were used as fixed effects in the univariate models. The results of this study indicated that the Bayesian ridge regression and ridge regression-best linear unbiased prediction methods were superior to other genomic prediction methods which were used in this study under terminal drought stress and well-watered conditions, respectively. Conclusions This study provided a robust, quick, and cost-effective machine learning-enabled image-phenotyping pipeline to improve the genomic prediction accuracy for days to maturity in wheat. The results encouraged the integration of phenomics and genomics in breeding programs.

A classic approach for determining genomic prediction accuracy under terminal drought stress and well-watered conditions in wheat landraces and cultivars

“…The ability of GS to enhance plant breeding is based on the fact that genotyping will soon become cheaper, and consequently, breeders will be able to save time and reduce phenotyping tasks [ 8 , 15 ]. Although, higher prediction accuracy will be obtained with increasing marker density [ 56 , 59 ], if the marke set is in a direct linkage with the traits of interest, GS would potentially explain all genotypic variables which are in LD [ 60 ].…”

“…Drought stress affects many vital biological processes in plants such as photosynthesis, respiration, and metabolism [ 6 , 7 ]. In the Persian plateau, where most of the climate zone is arid or semi-arid, farmers are very well-trained during the centuries to deposit rainwater throughout spring for irrigating farms at the end of the growing season when there is no rain during the seed development stage [ 8 ]. The Persian farmers irrigate their farms two to four more times with the stored water after spike appearance to avoid yield loss due to late-season drought stress [ 8 ].…”

“…Genomic selection (GS) produces a genomic estimated breeding value (GEBV) using all minor and major effects QTL in the genome, so that candidate genes can be selected by genotyping before phenotyping [ 15 ]. GP uses all markers within a model to train a prediction model in the training set (TS) which includes all genetic effects, without considering how minor the genetic effects are [ 8 , 15 ]. The model will be applied to a validation set (VS) to estimate the accuracy of GP.…”

The 4.(n) wheat breeding era: genomic-based predictions becoming part of the daily routine and beyond

“…One way phenomics in fact enters in GS is through adding secondary correlated traits in the models, as covariates or multivariates (e.g. Shabannejad et al, 2020). In this scenario, "special traits", such as those obtained through images and temperature measurements, gain special relevance.…”

“…Days to maturity was predicted three to The 4. (n) wheat breeding era 3-5 four times more accurately when drought tolerance ratios obtained from machine learning-treated images were added as covariates in the GS equation (Shabannejad et al, 2020). Obviously, each trait is impacted differently by the use of correlated traits in GS.…”