Within plant breeding new data generating techniques in the field of genotyping, phenotyping, omics and envirotyping present new opportunities for increasing the efficiency of breeding programmes. To take advantage of these developments we provide new statistical techniques and software.
A statistical genetics pipeline to help breeders make decisions
- A wide variety of experimental designs that allow precise estimation of genotypic effects and contrasts while correcting for noise factors.
- Quantitative genetic analyses for all types of breeding populations that deliver estimates for important statistical genetic parameters such as repeatabilities, heritabilities, and genetic and environmental variances, covariances, and correlations.
- Exploration and quantification of genetic diversity and distances using sequence, molecular marker and phenotypic data. Preparation for genotype-to-phenotype modelling.
- To locate quantitative trait loci on chromosomes, a powerful map construction algorithm is available that efficiently produces genetic maps for biparental and multiparental populations.
- Extensive suite of methods and visualizations to deal with the major problem in plant breeding and evolutionary biology: genotype by environment interaction, where genotypic differences are conditional on the environment, and achieving the breeders’ main objective, selection of the best genotypes, becomes complicated.
- A powerful mixed model based approach identifies quantitative trait loci (QTLs) for whichever kind of breeding population, mating type, ploidy level, number of traits and environments.
- Data fusion and integration of high throughput genotyping, phenotyping and omics data.
- Decision support for selection of superior genotypes and crossing partners.
- Visualization tools to support reporting and decision making.
A statistical genetics pipeline for innovation and integration
- Designed to allow smooth information exchange of data between the statistical algorithms, visualization tools, databases and applications
- Problem solving across large collections of high dimensional data • (genomic and phenotypic)
- The statistical genetic pipeline API provides a standard interface to access the tools and algorithms to serve analysis to their applications
- The pipeline embeds new algorithms and methods for spatial analysis of field trials by two-dimensional splines.
- The generation of conditional QTL genotype probabilities for all kinds of breeding populations by a continuous time Markov process, used in genetic map construction and QTL mapping, and mixed model estimation algorithms for multi-trait analysis.
- The pipeline is based on R procedures that call on asreml-R mixed model procedures.
- The pipeline makes state of the art statistical and decision support methods accessible to a wide audience of breeders and geneticists via an intuitive user interfaces. The modularity of the pipeline allows easy adaptation to specific user requirements.
The H3 Consortium:
Drawing on our complementary world-class expertise, tools and experience in plant breeding, analytics and visualization, on projects including: Generation Challenge Program, Integrated Breeding Platform, Breeding View statistical software and on various EU research programs such as EU-MABDE and EU-Whealbi: Biometris, VSNi and James Hutton have come together to build H3.
• High throughput phenotyping
• High throughput genotyping
• High throughput data analytics
Fred van Eeuwijk, Martin Boer, Chaozhi Zheng, Marcos Malosetti, Willem Kruijer, Bart-Jan van Rossum, Maikel Verouden, Ron Wehrens, Darren Murray, Allen Green, Jim Twynam, David Marshall, Iain Milne