Co-PIs: C. Thomas Hash (ICRISAT-Patancheru, Hyderabad, India), Stephen Kresovich (University of South Carolina, USA), Sharon Mitchell (Cornell University, USA), and Kassa Semagn (CIMMYT-Nairobi, Kenya)
Senior Personnel: Theresa Fulton (Cornell University, USA)
Advanced strategies of plant breeding are rapidly evolving from molecular breeding methods that manipulate a few major genes or genomic regions to complex, coordinated methods that incorporate simultaneous selection across many genomic regions (genome-wide or genomic selection). The future of plant molecular breeding (from conservation of key landraces to targeted hybrid deployment) must integrate and mine huge data sets and employ mathematical models that can predict which genotypes will perform well in specific environments. This information will then be used to help breeders and farmers identify the optimal parental combinations and select progeny genotypes (for a particular target environment) that most efficiently maximize agricultural output (yield and quality) with minimal use of constrained resources. This unique opportunity is founded on the pillars of inexpensive high marker density genotyping, high quality phenotyping, and appropriate mathematical models for prediction and experimental design. This research effort is planned to solve the technical problems associated with applying high-throughput high density genotyping and developing an operational "model" for a wide range of crops. The project will exploit and optimize next generation sequencing technologies as high-throughput genotyping assays for maize and sorghum varieties important to farmers in sub-Saharan Africa (SSA) and other semi-arid, tropical regions of the world. Using next generation sequencing technologies combined with high levels of multiplexing and very inexpensive sample preparation methods, the project will establish a platform to provide 100,000-1 million markers per individual for a cost substantially below current field and laboratory costs. This platform will produce robust protocols for the entire pipeline, from DNA preparation to bioinformatic analyses, which are essential for genomic selection. Moreover, these approaches will be generalized so in the future any large number of individuals of any species can be genotyped with equal ease. Reducing the genotyping costs and increasing its applicability to all crops will enable plant breeders to employ genomic selection to facilitate crop improvement.
Genomic selection has the potential to significantly increase the rate of gain in agricultural productivity per year in the developing world. In this coordinated project, efforts focus on solving key technology components necessary to give half the world's population access to advanced approaches for making better crop varieties specifically tailored to challenging environments and production systems. This project will make protocols available through the project website, sequencing reads through NCBI Short Read Archive, and diversity data through the Panzea and Gramene websites. Additionally, the project will provide training to researchers at Cornell and through two symposia and training modules held in Kenya and India.
DNA sequencing technologies have improved 20 million fold over the last two decades, and although these advances have changed all areas of biology, it has particularly powerful implications for plant breeding. The fusion of extensive genetic markers, field evaluations, and quantitative models now allow breeders to accurately predict crop field performance across several breeding generations based solely on genetic markers. Importantly, genetic predictions can be made fast, while field performance data can take years to collect. This is a fundamental shift in breeding that can accelerate crop improvement throughout the world’s breeding programs. This project focused on one building block of this revolution: reducing the cost of genotyping to below that of field evaluations, so breeders can cost effectively make decisions based on genetics. This project successfully developed the molecular biology protocols and open source bioinformatics for a robust cost-effective genotyping by sequencing platform. Specifically, a next generation sequencing method was developed that takes advantage of reduced representation to enable high throughput genotyping of large numbers of individuals at a large number of SNP markers. The resulting GBS protocol is relatively straightforward, robust, and cost-effective. Additionally, the project developed a bioinformatics pipeline, TASSEL-GBS, designed for the efficient processing of raw GBS sequence data into SNP genotypes. These approaches were first applied to the breeding efforts of maize and sorghum for Sub-Saharan Africa. Through training courses and outreach efforts the approaches have now been applied to over 300 species important for plant breeding, conservation, and basic biology.
