Predictive breeding for maize: Making use of molecular phenotypes, machine learning, and physiological crop models

Link to full Paper

Short Summary:

• The start of genomic prediction began with Meuwissen, Hayes, & Goddard 2011.
• Standard genomic selection models may benefit from additional information along the central dogma, or outside of it (i.e., expression, metabolites, microbiome).
• Incorporating crop growth models with genomic prediction may further enchance prediction models.

Abstract

Maize (Zea mays L.) has been a focus of scientific research and breeding for over a century. It is also one of the most economically important crops in the world, with a value of approximately US$50 billion per year in the United States alone. Additionally, maize has long been the model species of choice for the study and exploitation of hybrid vigor, and it continues to be one of the world’s most efficient converters of photosynthetic energy into starch. This review discusses the history and future of maize predictive breeding in the context of both genotype centric methods, and those focusing on genotype × environment × management interactions. Current prediction challenges are highlighted, as well as important advances in technology, methods, datasets, interdisciplinary collaborations, and scientific culture that will enable accelerated progress in predictive maize (and other crop species) breeding for years to come.

This paper is published in Crop Science (https://doi.org/10.1002/csc2.20052)