RESEARCH: Predicting Genotypic Variation in Growth and Yield under Abiotic Stress through Biophysical Process Modeling

由于世界人口的增加以及更可能的温度和干旱胁迫，对高质量粮食作物的需求不断增加，这需要通过育种计划进一步改进作物。这些计划的一个主要限制是理解遗传信息如何影响植物的特征，从而提高可食用部分的数量。此外，如果植物被置于低降雨量或高温等新的压力环境中，预测性理解就更少了。改善育种的一个可能的研究途径是更好地将储存在基因中的信息变成植物的特征，这些信息将植物的生物学特征，如光合作用速率或分配给可食用根的资源量，与物理世界相结合，如可用水量或过度热浪。目前建立这些联系的方式通常需要为每一种新的作物或环境收集新的数据，例如新的土壤、新的温度范围，甚至是显示植物分配给可食用根的资源量变化的新的改良线。这种对新信息的持续需求最终减缓了育种计划和植物科学家快速响应社会需求的能力。该项目将测试一种新的方法，该方法使用大量数据来计算特定植物特性在各种环境条件下由给定植物品系显示的概率。具体地说，该项目将通过在植物中发送电脉冲来持续测量植物的表现，将产生的数据与大型数据集相结合，这些数据集显示了哪些基因是活跃的，以及在任何给定时间对植物内主要代谢途径做出贡献的生物相关分子的水平。这一新方法需要高性能计算来多次测试作物表型改善的可能性。这些高性能计算方法将成为现代、具有竞争力的劳动力的核心部分。在这方面，该项目将为高中教师提供讲习班，让他们在课堂上使用高性能但开放源码的计算工具。此外，该项目将使用非常成功的威斯康星州快速学习系统(http://www.fastplants.org/).)为6-12年级的学生开发生物和定量学习的实验和计算模块随着世界人口的增加，迫切需要遗传技术来提高作物的生长、产量和对非生物胁迫的抵抗力。限制作物改良的速度是关于调节基因组和表型组之间关系的生物物理过程的一个关键知识缺口，阻碍了在新环境中预测新基因型的表型的能力。作为弥补这一差距的第一步，高通量表型和生物物理过程建模的组合将包括影响植物碳代谢、水力学和资源分配的关键基因的等位基因变异，所有这些基因都已知影响植物的抗旱性和耐热性。作物多样化过程中的不同选择压力导致了油菜作物之间广泛的表型差异，使其成为一个很好的研究系统，既可以将器官水平的措施连接到转录和代谢表型水平，也可以连接到产量水平，并测试预测过程模型。过程模型将利用连接细胞过程和最终整个植物生理与代谢产物和基因转录物等调节中间体的机械联系来开发和改进。如果成功，所开发的模型将能够在不同的基因类型和环境中预测全植物的胁迫反应表型。该项目的目标是：1)采用一种新的高通量和实时的表型方法来测量8个油菜亲本嵌套关联图谱(NAM)在干旱和热胁迫条件下的日生理动态；2)利用碳代谢、水力学和资源分配的生物物理过程模型来预测重组油菜自交系(RIL)群体的产量，以测试昼夜节律、转录、代谢和生理QTL之间的系统水平联系；3)使用AIM 2中使用的RIL种群测试生物物理过程模型在高温和干旱胁迫环境下的预测能力。本项目产生的所有数据和资源将通过Github项目和NCBI短读档案等长期开放存储库向公众开放。

Rising demand for high quality food crops due to increasing world populations along with more likely temperature and drought stress requires further crop improvements from breeding programs. A major limitation to these programs is an understanding of how the genetic information affects the characteristics of plants that improve the amount of edible portions. Moreover, the predictive understanding is even less if the plants are placed in new, stressful environments like low rainfall or high temperature. A likely avenue of inquiry to improve breeding is to better connect how information stored in genes becomes traits of plants that combine their biology, such as photosynthetic rate or the amount of resources allocated to an edible root, with the physical world, such as the amount of water available or an excessive heat wave. These connections are currently often made in a way that requires new data collection for every new crop plant or environment such as a new soil, new temperature range or even a new improved line that shows variation in the amount of resources the plant allocates to an edible root. This continuous need for new information ultimately slows down the breeding program and the ability of plant scientists to quickly respond to the needs of society. This project will test a new approach that uses large amounts of data to calculate the probability that a particular plant characteristic will be displayed by a given plant line under various environmental conditions. Specifically, the project will measure plant performance continuously by sending electrical pulses through plants, integrating the data generated with large data sets that show which genes are active as well as the level of biologically relevant molecules that contribute to major metabolic pathways within the plants at any given time. This new approach requires high performance computing to test many times how the probability of phenotypic improvement in the crop may occur. These high performance computing approaches will become a core part of a modern, competitive workforce. In this regard, the project will provide workshops for high school teachers in the use of high performance yet open source computing tools in their classrooms. In addition, the project will develop experimental and computational modules in biological and quantitative learning for students in grades 6-12 using the highly successful Wisconsin FastPlants system (http://www.fastplants.org/). With increasing world populations, genetic advances to improve crop growth, yield and resistance to abiotic stress are a pressing need. Limiting the speed of crop improvement is a crucial knowledge gap regarding biophysical processes that modulate the relationship between the genome and phenome, hindering the ability to predict the phenotype of novel genotypes in novel environments. As a first step towards bridging this gap, a combination of high-throughput phenotyping and biophysical process modeling will incorporate allelic variation at key genes affecting plant carbon metabolism, hydraulics, and resource allocation, all of which are known to impact drought- and heat-stress resistance in plants. Variable selective pressures during crop diversification have caused extensive phenotypic variation among B. rapa crops, making it an excellent study system to both connect organ-level measures both down to the level of transcriptomic and metabolomic phenotypes and up to yield and to test predictive process models. Process models will be developed and refined using the mechanistic links that connect cell processes and ultimately whole plant physiology to regulatory intermediates such as metabolites and gene transcripts. If successful, the models developed will enable prediction of whole-plant stress-response phenotypes in heterogeneous genotypes and environments. The goals of the project are to: 1) deploy a novel high-throughput and real-time phenotyping method to measure diel physiological dynamics in eight B. rapa parental Nested Association Mapping (NAM) lines under drought- and heat-stress conditions; 2) predict yield in a Recombinant Inbred Line (RIL) population of B. rapa using a biophysical process model of carbon metabolism, hydraulics and resource allocation to test systems-level links between circadian, transcriptomic, metabolomic, and physiological QTL; and 3) test the predictive ability of the biophysical process model under heat- and drought-stress environments using the RIL population used in Aim 2. All data and resources generated in this project will be made accessible to the public through long-term open access repositories such as Project Github and the NCBI Short Read Archive.