Using a novel PLS approach for envirotyping of barley phenology and adaptation

Abstract

Improving grain yield and adaptation is achieved by synchronising crop phenology with the environment. Phenology research is complex and encounters analytical challenges in characterising genotype x environment (GxE) interactions. This paper presents a simple approach that helps to explain the environmental drivers of phenology. Photoperiod and temperature are major environmental cues for expression of crop developmental genes, and sensitivity to photoperiod is thought to be the major cause of maturity differences among Australian spring barley varieties. However, temperature and photoperiod show similar seasonal trends and strong autocorrelation makes it difficult to distinguish their relative importance in crop development. Partial Least Squares regression (PLS) was developed to handle large data sets with many correlated explanatory variables and only one dependent variable. Across 35 environments encompassing a wide range of temperature and photoperiod regimes in southern Australia, a PLS model described more than 90 % of the phenotypic variation in time to anthesis of three adapted barley cultivars. The PLS outputs defined the critical periods when photoperiod and temperature were most influential, and revealed that temperature effects are of equal or greater importance than photoperiod in determining anthesis date, which is a new finding for genotypes adapted to Australian environments. Insight into the previously elusive differential responses to changes in daily average, maximum, and minimum temperatures will assist in developing flowering models for growers that are more accurate, and assist breeders in the genetic dissection of phenology for target environments.