The phenotype and the components of phenotypic variance of crop traits

Abstract

Field Crops Research focuses on both experimental and modelling research at the field, farm and landscape level on temperate and tropical crops and cropping systems, with a focus on crop ecology and physiology, agronomy, plant genetics and breeding. Emphasis is on species more relevant to ensuring global food security and on their performance in the field.

An increasing number of manuscripts dealing with genetics and breeding also deal with trait phenotype and individual components contributing to the total phenotypic variance. Some of these either fail to deliver new scientific insight, have experimental deficiencies or use inappropriate tools for analysis. A lack of explicit theoretical frameworks is common. The aims of this article are to identify research gaps and new scientific developments, highlight common experimental and analytical deficiencies, and outline editorial criteria addressing this specific area, where the journal would favour: (1) rigorous account and interpretation of the components of phenotypic variance, (2) closing the gap between phenotypic and genotypic characterisation of crop traits, (3) developing and applying advanced approaches for quantitative environmental characterisation; (4) linking with appropriate theories to improve genetic, agronomic, physiological and ecological interpretations of the phenotype and its drivers, and (5) methods for efficient screening of plant populations and segregating progenies for yield potential and stress adaptation, with an emphasis on biological mechanisms.

Introduction

Mahner and Kary (1997) identified at least five different notions of genome, seven of genotype, and five of phenotype. West-Eberhard (p. 31, 2003) discussed these concepts from the perspective of developmental biology, thus defining genome as the full complement of DNA in a cell, the phenotype as including all traits of an organism other than its genome, and the genotype as the genetic makeup by which an individual or one of its traits can be characterised in genetic comparisons with other individuals or their phenotypic traits. Although rarely used, the term phenome seeks symmetry with genome, and further emphasises that the phenotype is the individual outside the genome (Mahner and Kary, 1997). According to these definitions, wheat phenology, rice morphology and maize yield are part of the phenotype as are cell walls, enzymes, transcription factors, hormones and metabolic pathways; the profile of mRNA, an operational definition of gene expression, is also part of the phenotype (Nachtomy et al., 2007). This perspective highlights the relevance of phenotypic traits for agriculture and underscores the developmental loop: gene expression → phenotype (e.g. abscisic acid) → gene expression (e.g. Chandler and Robertson, 1994).

The prevailing model in crop science partitions the phenotypic variance into genetic (G), environmental (E) and interaction (G × E) components, plus a residual ‘error’ term. A variant of this model also accounts for maternal effects on trait expression (Galloway et al., 2009). The genetic component of the phenotypic variance of crop traits attracts most attention as reflected in the outstanding rate of scientific and technological progress in this area (Tuberosa et al., 2011, Edwards et al., 2013). The need for phenotyping techniques that keep pace with genomic approaches has been recognised (Edmeades et al., 2004, White et al., 2012, Rebetzke et al., 2013a). Yet despite these efforts, quantification of phenotypic traits relevant to potential yield and adaptation to abiotic stresses remains a bottleneck in plant breeding. Rigorous quantification of the environment is rare despite the widespread interest in the interaction between genotype and environment. In an agronomic context, the interaction is often extended to include management practices as a component of the environment, or explicitly as a third factor in the interaction (Messina et al., 2009).

Experimental design and statistical methods that deal with the components of the phenotypic variance and improve the correlation of phenotype with genotype are improving steadily, partially by reducing the error component (Crossa et al., 1991, Cullis et al., 2006, Gauch et al., 2011). However, a lack of explicit theoretical frameworks is common and biological and agronomic interpretations lag behind statistical descriptions (Edmeades et al., 2004).

An increasing number of manuscripts deal with trait phenotypes and the underlying components of the phenotypic variance; few formulate working hypotheses, some of them either fail to deliver new scientific insight, have experimental deficiencies or use inappropriate analytical tools. Owing to rapid progress in these fields, any attempt to be prescriptive in relation to methods is bound to be ephemeral. Instead, this article identifies gaps and new scientific developments to outline editorial criteria favoured by Field Crops Research in the area of phenotype and the components of the phenotypic variance. At the same time, the importance of basic requirements for adequate plot size, bordering, experimental design and the need for representative environments are reiterated.