GplusE: Genomic selection and Environment modelling for next generation wheat breeding

Lead Research Organisation: University of Edinburgh
Department Name: The Roslin Institute

Abstract

Despite its importance and growing demand within the UK, and globally, the rate of increase in wheat yields on UK farms have stagnated. To meet global future demand, annual wheat yield increases must grow to at least 1.4% and increasing the rate of genetic improvement using modern approaches is one way to do this. The ability to record vast quantities of genetic and phenotypic information cheaply (e.g. genetic markers and spectral images of field trials - termed in this proposal as Genomics and Phenomics) presents a new opportunity for increasing the rate of genetic improvement.
The rate of genetic improvement is affected by (1) the accuracy of selection, (2) breeding cycle time, (3) selection intensity, and (4) the amount of genetic diversity to be selected upon. In the medium to long term, concerns about genetic diversity are being addressed through national and international projects to introgress traits and alleles from landraces and progenitor species. However, the major barrier to the immediate increase in the rate of genetic improvement in wheat is the length of the breeding cycle time. Even at their fastest wheat breeding programs require at least four to six seasons to complete a cycle, principally due to the time required to reduce the number of individuals for selection to a subset that can be intensively phenotyped. Genomic selection (GS) is a new breeding tool that, amongst other advantages, can dramatically reduce breeding cycle time as selection can occur without the need to record phenotypes. In wheat this means breeding cycle time could be reduced to one season, dramatically increasing the rate of genetic improvement. In the extreme, using glasshouses to complete 2 cycles of selection per year, 10 cycles could be undertaken in the 5-year time frame currently taken for a single selection cycle.
GS uses a training population that is phenotyped and genotyped to construct a prediction equation. This equation is used to predict the breeding values of unphenotyped individuals, which, in wheat, would allow reduction of the breeding cycle to one season. GS assumes that saturating the genome of all individuals with molecular markers and estimating the effect of these markers (i.e. training the prediction equation) will allow capture of a large proportion of the genetic variation caused by the underlying quantitative trait loci. If the proportion of the captured genetic variation is large and well estimated the prediction equation will be able to make accurate predictions about breeding values. Similarly, in Phenomics the phenotype could be saturated with descriptors, which could lead to a better separation of its environmental and genetic components as well as generating more precise phenotypes.
Creation of training populations is a required investment for GS and strategic use of resources to achieve the required size is needed to optimize the cost and benefit of GS. Use of a genotyping and imputation strategy is paramount for reducing costs. Field trials are also costly. Use of novel high-dimensional approaches for capturing extra traits and variables (Phenomics) could enhance the value of field trials generally, as well as enabling more powerful GS.
This proposal will use field trials and simulation to design and evaluate Genomics and Phenomics strategies for increasing rates of genetic improvement in wheat. This will include GS training population designs and low cost collection of genotype data, assessment of the properties of high-dimensional environmental descriptors and quantification of their power, assessment of the properties of trait phenotypes collected by high-dimensional data recording devices and quantification of their relationships to standard traits. Results will be generalised to other species with breeding programs similar to those of wheat as well as to other type of experiments and field trials (e.g. National List evaluations).

Technical Summary

This proposal will develop and evaluate strategies for next generation wheat breeding based on the use of Genomic and Phenomic data, which can now be recorded in large quantities cheaply. It will bring "Big Data" to wheat breeding. Genomic selection (GS) could revolutionize wheat breeding: cycle times could be shortened; accuracy and intensity of selection maintained or increased; and more widespread selection for hard to measure traits could be undertaken. The key advantage of GS in comparison to traditional breeding is the ability to use high-dimension genetic marker information to make accurate predictions of genetic value for selection candidates without having to wait for phenotypes to be collected.
Analogous to GS, Phenomics uses instruments to record high dimensional phenotypic (e.g. spectral images) and environmental (e.g. soil electro-magnetic imaging) data. These may be determinants or indicators of complex trait phenotypes such as yield. More accurate modelling of environment and phenotype could enable faster genetic improvement in wheat, or increased efficiency of assessment in any field trial. Phenomics is highly valuable for wheat breeding. This project will scale that value to field level, enabling it to drive genetic improvement.
This project will use a large field trial combined with simulation to study the key factors in the application of GS to wheat breeding: training population design and genotyping strategy. It will integrate GS with Phenomics so that they empower each other.

Planned Impact

Genomic selection and Environment modelling for next generation wheat breeding (GplusE) links Phenomics and Genomics to deliver to the wheat breeding community a platform to greatly increase the rate of progress possible through breeding. Wheat is the UK's major crop and has the 3rd largest production of any cereal globally. This project has the potential to benefit individuals and organisations worldwide for whom improvement in wheat yields is important. This ranges from farmers in the developing world, through millers and bakers, to anyone buying bread in their local supermarket. The impact of the methods we develop will be seen first by the UK wheat breeders involved in this project, with delivery of improved varieties to market in the following few years.
More specifically:
The commercial breeding partners will benefit immediately by:
1. Initiation of propriety training populations of direct relevance to their breeding programs;
2. Protocols for cost reduction of genotyping by use of imputation and of phenotyping by remote capture of covariates;
3. Access to data generated in the project and to source code for programs required to implement GplusE.
Other breeders will benefit by publication of protocols for the application of GplusE within their own breeding programmes, including access to compiled versions of software developed within the project. This will be available following their publication.
Agronomists and field crop researchers will benefit from exemplars of the use of Phenomics and precision agriculture to improve the accuracy of treatment comparisons in field trials, for example in trials comparing agronomic inputs.
Suppliers of precision agriculture services and genotypes will benefit from new market opportunities in supplying services to field trials operators, including breeders.
The academic crop research community will benefit from improved techniques for field trials and demonstration of how novel physiological and other traits could be incorporated into breeding programmes within a quantitative genetics framework.
Longer term (>8 years) the linking of Phenomics and Genomics may result in development of varieties more adapted to specific environment conditions allowing, for example, automated switching of varieties during drilling to match micro-environmental conditions. This benefits the farmer in improving yield and the environment by reducing inputs currently required to compensate for variety weaknesses.
NIAB / RI will benefit from exposure to complementary expertise in each other's institutes. For NIAB this will result in improved application of big-data methods to their research interests in agricultural science and genetics. For RI, it will open opportunities for broader application of its quantitative genetics research and future funding and collaboration with both private and public sector crop and plant science communities.
The PDRAs funded by this project will benefit from contact with each other and with staff outside their host institutes. They will gain knowledge and expertise beyond the confines of their project areas.
This project will genotype, phenotype and analyse 3,000 lines. The size of the project and the disciplines involved, encompassing quantitative genetics, molecular genetics, plant physiology, Phenomics and agronomy will provide opportunities for training of scientific and technical staff in these fields. The research will also generate opportunities for field visits and workshops to stakeholders from outside RI and NIAB. NIAB Innovation Farm showcases innovation in agriculture, provides free support and assistance for small to medium businesses in the East of England and hosts national and international workshops. This infrastructure will be used to inform a wider public and scientific community.

Publications

10 25 50
 
Description 1. We have developed novel strategies for the utilisation of genomic selection in pre-breeding programs in crops (Gorjanc et al., 2015) and within crop breeding programs themselves (Gaynor et al., accepted), for the efficient generation of genomic data in crops (Hickey et al., 2015; Gorjanc et al., 2017; Gorjanc et al (accepted)) and for the design of genomic selection training sets in crops (Hickey et al., 2014).
2. The principles that we had previously established on the effective design of genomic selection training sets apply to crops.
3. The simulations of our strategy for the utilisation of genomic selection in crop breeding programs show that if it was used crop breeding programs could increase the rate of genetic gain three fold.
4. We have been able to extend our simulation program AlphaSim (for simulating sequence, genotypic, phenotypic, and pedigree data) to a range of crop scenarios, adding several new features that apply to crops, allowing breeding programs to be simulated rapidly performing sensitivity analyses on key parameters of interest. We have used it to assess genotyping strategies in crop breeding programs (Gorjanc et al 2016). Given the potential of AlphaSim, and its increased use by the academic community, and we have recently developed a Graphical User Interface (AlphaSim GUI) that will make the use of AlphaSim much more accessible to a wide range of users.
Exploitation Route The first steps are already in place building on the industrial partnership award. We are aware that this technology also has interest, ultimately, for pharmaceutical companies and other industrial biotechnology companies.
Sectors Agriculture, Food and Drink,Manufacturing, including Industrial Biotechology

 
Description They have contributed to changing the approach to plant breeding, accelearting change and introducing new genomic technology into the modus operandi of plant breeding companies.
First Year Of Impact 2016
Sector Agriculture, Food and Drink,Manufacturing, including Industrial Biotechology,Pharmaceuticals and Medical Biotechnology
Impact Types Societal,Economic

 
Description Newton Fund Workshop Brazil
Amount £52,000 (GBP)
Funding ID 228949780 
Organisation British Council 
Sector Charity/Non Profit
Country United Kingdom
Start 04/2016 
End 09/2016
 
Description Newton Fund Workshop Mexico
Amount £37,550 (GBP)
Funding ID 2016-RLWK7-10399 
Organisation British Council 
Sector Charity/Non Profit
Country United Kingdom
Start 04/2017 
End 03/2018
 
Description GplusE consortium 
Organisation Elsoms Seeds
Country United Kingdom 
Sector Private 
PI Contribution Execution of field trials Collecting phenotype data Creating and processing of genotype data Data analysis
Collaborator Contribution Created test crosses Supplied seeds
Impact Journal publications (partially still in preparation) Genotype data sets Phenomics data sets
Start Year 2015
 
Description GplusE consortium 
Organisation KWS UK
Country United Kingdom 
Sector Private 
PI Contribution Execution of field trials Collecting phenotype data Creating and processing of genotype data Data analysis
Collaborator Contribution Created test crosses Supplied seeds
Impact Journal publications (partially still in preparation) Genotype data sets Phenomics data sets
Start Year 2015
 
Description GplusE consortium 
Organisation Limagrain
Country France 
Sector Private 
PI Contribution Execution of field trials Collecting phenotype data Creating and processing of genotype data Data analysis
Collaborator Contribution Created test crosses Supplied seeds
Impact Journal publications (partially still in preparation) Genotype data sets Phenomics data sets
Start Year 2015
 
Description GplusE consortium 
Organisation RAGT Seeds
Country United Kingdom 
Sector Learned Society 
PI Contribution Execution of field trials Collecting phenotype data Creating and processing of genotype data Data analysis
Collaborator Contribution Created test crosses Supplied seeds
Impact Journal publications (partially still in preparation) Genotype data sets Phenomics data sets
Start Year 2015
 
Title AlphaPlantImpute 
Description AlphaPlantImpute is a software package designed for phasing and imputing genotype data in plant breeding populations. AlphaPlantImpute can be implemented within and across bi-parental populations to phase and impute focal individuals genotyped at low-density to high-density. 
Type Of Technology Software 
Year Produced 2018 
Impact This package was found to be extremely useful by our project partner global breeder KWS Saat SE. 
URL https://alphagenes.roslin.ed.ac.uk/wp/software/alphaplantimpute/
 
Title AlphaSim 
Description One of the fundamental questions in populations dynamic is assessing how changes in the current structure and environment affect the structure composition in both the short and long-term. Plant and animal breeding programs benefits from having a tool to evaluate the potential of different selection strategies or new emerging technologies to improve population performance. Empirical datasets to assess the effect of different factors on one population are difficult to collect, since they require substantial financial and time investments and are subject to noise and error. Simulation is a key tool for both researchers and breeders to assess the impact of different factors given a known historical and current population structure prior to implementation within a real-life setting. AlphaSim is a fast and flexible software tool that enables researchers and breeders to do this. Unlike other simulation tools, AlphaSim has the functionality to manipulate fine details of the population structure in order to simulate realistic scenarios and provides detailed outputs for use in downstream analyses. 
Type Of Technology Software 
Year Produced 2016 
Impact AlphaSim is a freely available software package that simulates genetic population and can assess breeding programs. The AlphaSim package includes a manual, tutorial, and access to technical support with the aim of benefiting the academic research community in animal breeding. This software package has already attracted users from a number of different academic institutions and has supported a number of peer-reviewed academic publications. These publications include: Potential of gene drives with genome editing to increase genetic gain in livestock breeding programs. 2017. Gonen, S, J. Jenko, G. Gorjanc, A.J. Mileham, C.B.A. Whitelaw, J.M. Hickey. Genetics Selection Evolution, 49:3. AlphaSim: Software for Breeding Program Simulation. 2016. Faux A. M., G. Gorjanc, R. C. Gaynor , M. Battagin, S. M. Edwards, D. L. Wilson, Sarah J. Hearne, S. Gonen, and J. M. Hickey. The Plant Genome vol. 9, no.3. AlphaSim is not only used in academic research, but has also attracted industrial collaborations. One such example is our recently awarded Innovate UK grant in collaboration with Driscoll's. 
URL http://www.alphagenes.roslin.ed.ac.uk/alphasuite-softwares/alphasim/
 
Title AlphaSimR 
Description AlphaSimR is a next generation software package in the line of our successful earlier package AlphaSim. The new package is accessible in a user-friendly way via an interface in the public domain environment R. The package is used for stochastic simulations of breeding programs to the level of DNA sequence for every individual. Contained is a wide range of functions for modeling common tasks in a breeding program, such as selection and crossing. These functions allow for constructing simulations of highly complex plant and animal breeding programs via scripting in the R software environment. Such simulations can be used to evaluate overall breeding program performance and conduct research into breeding program design, such as implementation of genomic selection. Included is the 'Markovian Coalescent Simulator' ('MaCS') for fast simulation of biallelic sequences according to a population demographic history [Chen et al. (2009)]. 
Type Of Technology Software 
Year Produced 2018 
Impact This package has rapidly expanded our possibilities to apply breeding simulation in research projects, both in academic research projects and for the breeding industry (most notably Driscolles and Bayer). Several graduate students used the package for their internship projects. 
URL https://alphagenes.roslin.ed.ac.uk/wp/software/alphasimr/
 
Description Big Data in Agriculture, Part of the DuPont Pioneer Plant Sciences Symposia Series, at Roslin Institute, 14-15 May 2018 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Other audiences
Results and Impact Symposium held at the Roslin institute, organised by members of my group, sponsored by third parties from the breeding industry
Year(s) Of Engagement Activity 2018
 
Description John Hickey Guest in Farming Today (BBC Radio 4) 
Form Of Engagement Activity A broadcast e.g. TV/radio/film/podcast (other than news/press)
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Public/other audiences
Results and Impact On Monday 26th September, The BBC Radio 4 Farming Today had Professor John Hickey as specialist scientist on the subject of breeding programs and scientific impact.
Year(s) Of Engagement Activity 2016
URL http://www.bbc.co.uk/programmes/b07w5xxq
 
Description Modern plant and animal applied genomics driven by genotype and sequence data, Universitat Politècnica de Valencia, 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Postgraduate students
Results and Impact Visiting teaching activity with advanced course in plant and animal breeding.
Year(s) Of Engagement Activity 2018
 
Description Modern plant and animal applied genomics driven by genotype and sequence data, University of Zagreb, Croatia, 17-19 July 2018 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Postgraduate students
Results and Impact Workshop organised and given by me and two other members of my group.
Year(s) Of Engagement Activity 2018
 
Description Researcher Links workshop at CNRG, INIFAP, Tepatitlán and Guadalajara, Mexico, 3-7 February 2018 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Postgraduate students
Results and Impact Workshop organised and given by me and the members of my group
Year(s) Of Engagement Activity 2018
 
Description Teaching course: Next Generation Plant and Animal Breeding Programs, Animal Science Department, University of Nebraska, Lincoln. 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Postgraduate students
Results and Impact Series of the lectures and workshops on Plant and Animal Breeding Programs exploring current practices and future areas
of research. The course was designed and imparted by John Hickey and key members of his team.
Year(s) Of Engagement Activity 2016
URL http://animalscience.unl.edu/next-generation-plant-and-animal-breeding-programs
 
Description The Expert Working Group on Wheat Breeding Methods and Strategies 
Form Of Engagement Activity A formal working group, expert panel or dialogue
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Professional Practitioners
Results and Impact Expert Working Group on Wheat Breeding Methods and Strategies seeks to exchange breeding methods research information and germ plasm to expert build capacity and support in wheat breeding programs, with more efficient breeding methods consistent with the latest scientific advances. The EWG is working on activities such us workshops, training courses, communications, and sharing of germplasm and information to reach larger pool of wheat breeders and trained in state-­of-­the-­art breeding methods.
Year(s) Of Engagement Activity 2015,2016,2017
URL http://www.wheatinitiative.org/activities/expert-working-groups/wheat-breeding-methods-and-strategie...