IPC Public:Community Portal
From IPC_Public
[edit] == The Maize Genetics Executive Committee ==
representing the maize community, suggests that an initial iPlant project be integrating phenotypic data from the metabolic pathway to cell properties to tissue to organ to organism to outdoor populations of individual inbreds/RILs/verified progenies, and finally to the diversity of maize landraces. We have much of the data and propose that the iPlant Collaborative would be an ideal environment to integrate, interpret, and then interrogate the data to make predictions about the functions of genes across these levels and to model the "systems" that make maize growth and reproduction robust despite the highly fluid nature of the genome (differences in gene content and gene number between lines). Pairs of maize inbreds are more divergent than human to chimp, and detailed QTL analysis is now uncovering ~20-30 loci controlling key plant properties such as days to flowering, height, ear size, etc. How can these whole plant characters be connected to the smaller scale pathways/cell and up to the larger scale of plant x environment and genetic structure of populations? 13 March 2008 Virginia Walbot
[edit] Across species, can we develop a robust approach to predict ecotypic differences in phenology as expressed in natural (field) environments?
Phenology (timing of key developmental stages) is key in determining adaptation of plant species whether in natural or managed ecosystems. Functional genomics is providing invaluable information on the basic clock mechanisms that regulate phenology, including the genetic basis of ecotypic differences. For some species, we are not too far from being able to accurately predict phenotypes from genotypes -- this latter being the ultimate "grand challenge." We lack, however, a coherent, integrated approach that is robust across plant species. This challenge is especially interesting because large component blocks of tools and data are available:
- Genomics database have extensive information on genetic networks involved in photoperiod response and floral development.
- Crop genetic/genomics databases (e.g., GrainGenes) have extensive data on major loci related to differences in photoperiod and vernalization responses of crop ecotypes.
- Pedigree databases (e.g., IRIS for rice and IWIS for wheat) hold vast amounts of information on genetic relations among ecotypes (cultivars or lines).
- Vast amounts of phenotypic data originating from multi-environment trials (METs) are held in the same or similar databases.
- The environmental data, most notably daily temperature records, are increasingly available through on-line sources.
- Ecophysiological models already predict phenology with an accuracy that is economically useful. For example, the IPCC AR3 and AR4 reports relating to agriculture largely base projections for impacts of climate change on outputs of ecophysiological models.
Advancing our ability to predict phenology would have demonstrable economic benefit on a 5 to 10 year time scale. For example, I recently attended a meeting of Arizona sorghum producers who wanted to know when and where to plant modern sorghum hybrids. In early plantings (e.g., March), they need hybrids that mature before mid-summer high temperatures limit yields; in late plantings, the hybrids must mature before low temperatures/high humidity slow grain filling and drying. At least four maturity loci are known in sorghum, but we know little about how these relate to the underlying physiology. Furthermore, we lack tools for rapid characterization of hybrids. The same ability to predict phenotypes from genotypes would allow us to vastly improve our predictions of impacts of climate change on sorghum production. Similar arguments would apply to essentially any major crop (annual, perennial or tree) and readily extend to species in natural ecosystems.--Jeff White 07:50, 11 March 2008 (MDT)
[edit] *Integrating New Model Systems
[edit] Integrated Information Management -- Crop Improvement on the Verge of a Revolution
Plant breeding is sometimes characterized as a “numbers game”, implying that a dominating element of chance or probability is best managed through large numbers of parents, crosses and selections. While chance and probability play an undeniable role in crop improvement (not unlike BLAST searches in genomics), modern plant breeding programs are highly data-intensive and rigorous analysis is emphasized to ensure efficient progress in the face of the uncertainties inherent in manipulating plant genetics and observing phenotypes in highly stochastic environments. The “numbers game” still exists, but success is vastly enhanced with efficient use of data for phenotypes, environments, and genetic loci. These efficiencies are being achieved in many major private sector improvement programs but are kept as proprietary resources. The public sector and smaller private sector programs individually lack the resources to develop the requisite data acquisition and management systems and then link their systems to the software tools that will allow them to maximize the information benefit, such as have been with the large and spectacularly useful genomics databases and informatics tools. Admittedly, data management has long concerned crop researchers, but various forces are converging to enable a revolution in how crop improvement is conducted, data are shared and mined, and the rapidity with which breeding impacts agricultural production. Trends creating this unprecedented opportunity include:
1. High throughput genotyping data is attaining costs per data point that are sufficiently low to permit marker based selection and breeding in segregating populations.
2. Database software and computers have the requisite power to manage the volumes of data that crop improvement and genotyping programs can generate, and need to access.
3. Where data are held by different institutions, the Internet is enabling seamless, nearly instantaneous data exchange.
4. Statistical analyses, largely supported by REML methods applied to mixed models, allow much better estimation of means and variances.
5. Improved analytic methods, notably association mapping, allow us to extract much more useful information from data on selections and lines and will allow cross-generational data mining.
6. Genomics and allied fields now provide various powerful tools for locating, sequencing and characterizing specific loci. We anticipate an information flood for new loci in coming years, although this flood is poised to demolish the “one-gene, one enzyme” paradigm: microRNAs, epigenetics and alternative pretranslational RNA splicing all complicate interpretation of sequence data in terms of specific gene functions.
7. Geospatial tools and ecophysiological models are reliable and flexible enough to assist interpretation of how the environment interacts with genotypes and crop management to produce phenotypes.
The question remains of how the crop improvement community can best capitalize on these trends and how the public sector can best develop the shared resources to accomplish for phenotypic and breeding records what similar collaborations have and continue to do for genomic research. The recently funded iPlant program (iplantcollaborative.org), hosted at the Univ. of Arizona presents a key opportunity to massively extend the reach of genomics and bioinformatics into plant improvement. However to participate, plant breeders need to communicate a coherent vision of how a major integrative effort will solve what iPlant terms “specific, compelling, and tractable Grand Challenges in the plant sciences.”
CHALLENGE TO THE CROP IMPROVEMENT COMMUNITY: We propose that the crop improvement community propose as a Grand Challenge the quantitative prediction of phenotypes on the basis of genetic, phenotypic, environmental and management data, translational genomics, and bioinformatics; and that the test-bed for this effort be ongoing crop improvement efforts in the public and private sectors.
GRAND CHALLENGE: We propose that iPlant, as a grand challenge, develop the public phenotypic and genomics databases and bioinformatic tools needed to support the broadest applications of genomics for plant improvement, thus advancing translation genomics into application genomics. The test-bed for this iPlant effort will be ongoing crop improvement efforts in the public and private sectors in all plants who are willing to share their breeding and pedigree records, genomic and phenotypic data, and knowledge for the public good. Submitted by: Jeff White, USDA ARS, Steve Baenziger, Univ. Nebraska
