Multi-environment QTL Mapping - QTLxE

GxE and QTLxE analysis

The purpose of statistical modeling in modern plant breeding is to predict phenotypic expression for multiple traits, across a range of environmental conditions, over developmental time, from molecular marker variation, genomic information and environmental inputs, for various types of (offspring) populations. This presentation tackles modeling GxE and its genetic basis - QTLxE.

Practical multi-environment QTL mapping (QTLxE)

Plant breeders usually evaluate their materials in several environments. This is because the relative performance of genotypes can change between environments, a well-known phenomenon called genotype by environment interaction (GEI). A QTL analysis that uses information of multiple environments allows onvestigation of the genetic causes underpinning GEI, i.e. QTL by environment interaction. Appropriate modelling of the genetic variance-covariance (VCOV) in the data os pf great importance when combining information from different environments in QTL analysis, amd mixed models are particularly suited for thisThe modelling of the VCOV is essential to produce appropriate tests of QTL effects. This practical illustrates a multiple environment QTL analysis based on mixed models. Special attention is placed on the appropriate modelling of the VCOV matrix. GenStat 14 is used to illustrate how to fit the models.

Training Materials - Mixed Model QTL Detection

This material introduces the different aspects of QTL detection with examples. It follows a didactical structure, first introducing basic concepts and gradually incorporating more advanced topics. Each section of the material has a short introduction followed by a number of questions around an example data set. The questions increase in complexity towards the end of the particular section. This is a hands-on training resource, where an example data set is presented and typical research questions are proposed to be solved by the researcher/trainee. An answer sheet with a short discussion of the different questions is included, making it possible to use the material for self-study.

The material has six examples/exercises: Exercise 1 – a simulated data set based on a 1 QTL model, to illustrate the basic principles of QTL detection using segregating populations (marker-based detection, simple interval mapping); Exercise 2 – a simulated data set where several QTLs are included in the model, to illustrate further developments in QTL mapping using segregating populations (eg: use of cofactors in composite interval mapping, QTL model selection, etc); Exercise 3 – a real data set where the principles learnt in exercises 1 and 2 are applied to a real situation; Exercise 4 – a real data set of several field trials, where the principles of field data analysis are discussed (mixed models for field experiments such as RCBD, Alpha design, spatial design, etc), estimation of important genetic parameters such as h2, genetic variance components, as well as the estimation of adjusted means; Exercise 5 – a real multi-environment data set, where phenotypic genotype by environment interaction (GxE) analysis is discussed as a first step of a QTL analysis. In this exercise, molecular marker information is included to discuss the principles of QTL mapping in multiple environments for detection of QTL and QTLxE; Exercise 6 – a real association mapping panel is used to illustrate the principles and statistical methods/models for linkage disequilibrium mapping.

These resources are downloadable below, together with an example data set and relevant references.


Back to top