Utility of climatic information via combining ability models to improve genomic prediction for yield within the genomes to fields maize project

Authors: JARQUIN, DIEGO - University Of Nebraska; DE LEON, NATALIA - University Of Wisconsin; ROMAY, MARIA CINTA - Cornell University; BOHN, MARTIN - University Of Illinois; Buckler, Edward - Ed; CIAMPITTI, IGNACIO - Kansas State University; Edwards, Jode; ERTL, DAVID - Iowa Corn Promotion Board; Flint-Garcia, Sherry; Gore, Michael; GRAHAM, CHRISTOPHER - South Dakota State University; HIRSCH, CANDICE - University Of Minnesota; Holland, Jim - Jim; HOOKER, DAVE - University Of Guelph; KAEPPLER, SHAWN - University Of Wisconsin; Knoll, Joseph - Joe; LEE, ELIZABETH - University Of Guelph; LAWRENCE-DILL, CAROLYN - Iowa State University; LYNCH, JONATHAN - Pennsylvania State University; MOOSE, STEPHEN - University Of Illinois; MURRAY, SETH - Texas A&M University; NELSON, REBECCA - Cornell University; ROCHEFORD, TORBERT - Purdue University; SCHNABLE, JAMES - University Of Nebraska; SCHNABLE, PATRICK - Iowa State University; SMITH, MARGARET - Cornell University; SPRINGER, NATHAN - University Of Minnesota; THOMISON, PETER - The Ohio State University; TUINSTRA, MITCH - Purdue University; WISSER, RANDALL - University Of Delaware; XU, WENWEI - Texas A&M University; LORENZ, AARON - University Of Minnesota

Submitted to: Frontiers in Genetics
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 12/21/2020
Publication Date: 3/8/2021
Citation: Jarquin, D., De Leon, N., Romay, M., Bohn, M., Buckler IV, E.S., Ciampitti, I., Edwards, J.W., Ertl, D., Flint Garcia, S.A., Gore, M.A., Graham, C., Hirsch, C.N., Holland, J.B., Hooker, D., Kaeppler, S.M., Knoll, J.E., Lee, E.S., Lawrence-Dill, C.J., Lynch, J.P., Moose, S.P., Murray, S.C., Nelson, R., Rocheford, T., Schnable, J.C., Schnable, P.S., Smith, M., Springer, N., Thomison, P., Tuinstra, M., Wisser, R.J., Xu, W., Lorenz, A. 2021. Utility of climatic information via combining ability models to improve genomic prediction for yield within the genomes to fields maize project. Frontiers in Genetics. 11:592769. https://doi.org/10.3389/fgene.2020.592769.
DOI: https://doi.org/10.3389/fgene.2020.592769

Interpretive Summary: Using genomic prediction methods to breed desirable characteristics in plants is an efficient alternative to traditional phenotyping selection methods. Because collecting weather data is becoming more automated and is consequently measured continuously, weather data could theoretically be used to predict plant development as part of other genomic prediction methods. This study measured the usefulness of including weather data variables in genomic prediction methods. Results showed that including real-time, weather data variables did not improve genomic prediction over the traditional method of using models to predict the effects of weather. However, more research is warranted on the link between observed weather and predicting its effects on plant development.

Technical Abstract: Genomic prediction methods provide an efficient alternative to conventional phenotyping selection for developing improved cultivars with desirable characteristics. New and improved methods are continually developed. Some of these attempts deal with the integration of different data types beyond genomic and/or pedigree information. Modern and automated systems offer the opportunity to capture weather data, practically, in a continuous manner over a time. In principle, this information could better characterize training and target environments to enhance predictive ability by incorporating weather characteristics as part of the genotype-by-environment (GxE) interaction component in the prediction models. Using a naive environmental kinship model across 30 environments comprising the Genomes to Fields (G2F) initiative in 2014 and 2015, we assessed the usefulness of including weather data variables in genomic prediction models in four different prediction scenarios that included predicting i) tested genotypes in observed environments; ii) untested genotypes in observed environments; iii) tested genotypes in unobserved environments; and iv) untested genotypes in unobserved environments. A set of 1,481 unique hybrids were evaluated for grain yield. Evaluations were conducted using five different models including main effects of environments (E), inbred markers for paternal and maternal effects GC (GP1, GP2); and the interactions between paternal and maternal marker effects SCA (GP1× P2 = GP1 × GP2), between inbred markers and environments (GP1 × E, GP2 × E and GP1 × P2 × E) and between the inbred markers and the Environmental Covariates ECs (GP1 × W, GP2 × W and GP1 × P2 × W). Results showed that the environmental kinship model did not improve predictive ability over prediction models that simply modeled the GxE interaction term. These results suggest that weather data per se is not enough for improving predictive ability. More research should be conducted to link the observed weather conditions with important physiological aspects in plant development to improve predictive ability through the inclusion of weather data.