Can traits measured from images and a few markers replace slower, destructive measurement?
Breeding populations are large, and traits such as biomass change through the season. Image traits can estimate seed weight, biomass and nitrogen use efficiency with less manual or destructive measurement, and a small SSR marker set can separate germplasm and follow the genes behind sweetness. Where dense genotyping is not yet available, as in hexaploid sweetpotato, genomic prediction is being prepared step by step.
Image-based phenotyping
Image traits from RGB, infrared and near-infrared sensors track biomass and nitrogen use efficiency through the season, in controlled environments and in the field.
RGB images of seeds, and RGB, infrared and near-infrared images of whole plants, are turned into image traits that stand in for the trait a breeder would otherwise measure by hand, in soybean, rice and wheat.
2023 · Agriculture
One image ratio predicts rice biomass at two nitrogen levels
Of 67 image traits from RGB, IR and NIR sensors, the ratio of projected shoot area to NIR intensity was a new predictor of rice biomass; a Bayesian regularized neural network reached R² 0.96 and 0.95 for fresh and dry weight.
2023 · Frontiers in Plant Science
Image models for wheat biomass give similar accuracy in two years
With 77 image traits and 16 machine-learning models, wheat fresh weight, dry weight and shoot area were estimated with about 90% accuracy; models fitted separately in two consecutive experiments gave similar accuracy, and Bayesian LASSO performed consistently well across traits and experiments.
2025 · Plant Stress
Review: high-throughput phenotyping for nitrogen-stress tolerance in rice
A review of how imaging platforms measure the dynamic component traits of nitrogen use efficiency, in controlled environments and in the field, and link them to genes and QTLs.
Markers and genomic prediction
A small SSR marker set can tag the genes behind sweetness and tell germplasm accessions apart, before any dense genotyping exists.
SSR markers are used in sweet corn for the sh2 and su1 sweetness genes and in sweetpotato for germplasm identity; for sweetpotato, a dosage-aware genomic-selection pipeline has been written and tested on simulated data, ready for the first real genotyping. Software →
2023 · Vietnam Journal of Agricultural Sciences
SSR markers linked to the sweetness genes sh2 and su1
Markers linked to the sh2 and su1 genes were identified in sweet corn lines, for use in selecting parents by marker rather than by test cross alone.