PLS0063 Biometrical methods for analyzing plant breeding trial data in the omics era, 3.5 Credits
Pass / Failed
The requirements for attaining different grades are described in the course assessment criteria which are contained in a supplement to the course syllabus. Current information on assessment criteria shall be made available at the start of the course.
Basic knowledge about genetics, quantitative genetics, linear regression models, analysis of variance and basic experience with R. At least a MSc degree in agronomy, biology, forestry, genetics, horticulture or enrolled as a PhD student.
Train Post-graduate students on biometrical and quantitative genetics to be applied in designing and analysing data from field trials, particularly variety testing by learning basic statistical models and methods for dissecting genotype-by-environment interaction considering advances made in the omics era.
At the end of this intensive 1-week course, the participant will:
- Refresh some basic ideas of Biometrical Methods for Plant Breeding Trial Data.
- Understand the conceptual framework of experimental design, multi-environment testing, use of association genetics and genomic prediction for plant breeding.
- Know how to design, analyze and handling genotype-by-environment interactions.
- Learn results that clearly show that biometrics works for plant breeding.
- Tell how to prepare phenotypic and genotypic data for analyzing plant breeding data with R.
- Run some R codes for analyzing plant breeding data involving different statistical models.
- Test if there are differences between breeding lines or populations, and cultivars.
1. Revisiting field plot techniques for variety testing design and analysis: principles of good design, blocking, spatial analysis
2. Brief overview of mixed models (BLUE, BLUP, REML, variance components)
3. Linear and other models for analyzing genotype-by-environment (G×E) interactions
4. Association genetics (GWAS) and genomic estimated breeding values (GEBV) for genetic enhancement
5. G×E in association mapping (GWAS) and genomic prediction (GEBV). Genomic prediction models for G×E (reaction norm models)
1. A brief introduction to R
2. Spatial analysis, use of incomplete block (a-lattice) and augmented design
3. Analysis of variance across sites and over years
4. Modeling multi-environment data under significant G×E
5. Including G×E in GWAS and GEBV
Formats and requirements for examination
Completion of all course assignments (pre-course readings, post-course practices-report), attendance and active participation in all theory lectures and computer practices, and oral exam to be taken with course leader.
Schedule during the intensive week:
8:30AM – 12:30AM Theory and Results from real data applications
1:30PM-5:30PM Practices with computer (each student brings theirs) – hands on for running R codes fitting a variety of stats models.