==== Front In Silico Plants In Silico Plants ISP In Silico Plants 2517-5025 Oxford University Press ISP-01-01-diy003 10.1093/insilicoplants/diy003 Review Systems models, phenomics and genomics: three pillars for developing high-yielding photosynthetically efficient crops Chang Tian-Gen 1 Chang Shuoqi 2 Song Qing-Feng 1 Perveen Shahnaz 1 Zhu Xin-Guang 1* 1 National Key Laboratory of Plant Molecular Genetics, CAS Center for Excellence in Molecular Plant Sciences, Shanghai Institute of Plant Physiology and Ecology, Chinese Academy of Sciences, Shanghai 200031, China 2 State Key Laboratory of Hybrid Rice, HHRRC, Changsha 410125, China * Corresponding author’s e-mail address: zhuxg@sippe.ac.cn 25 4 2019 2019 25 4 2019 1 1 10.1093/insilicoplants/diy00323 10 2018 17 12 2018 13 2 2019 © The Author(s) 2019. Published by Oxford University Press on behalf of the Annals of Botany Company2019This is an Open Access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited.Recent years witnessed a stagnation in yield enhancement in major staple crops, which leads plant biologists and breeders to focus on an urgent challenge to dramatically increase crop yield to meet the growing food demand. Systems models have started to show their capacity in guiding crops improvement for greater biomass and grain yield production. Here we argue that systems models, phenomics and genomics combined are three pillars for the future breeding for high-yielding photosynthetically efficient crops (HYPEC). Briefly, systems models can be used to guide identification of breeding targets for a particular cultivar and define optimal physiological and architectural parameters for a particular crop to achieve high yield under defined environments. Phenomics can support collection of architectural, physiological, biochemical and molecular parameters in a high-throughput manner, which can be used to support both model validation and model parameterization. Genomic techniques can be used to accelerate crop breeding by enabling more efficient mapping between genotypic and phenotypic variation, and guide genome engineering or editing for model-designed traits. In this paper, we elaborate on these roles and how they can work synergistically to support future HYPEC breeding. Crop yieldgenomicsHYPECmodel-guided breedingphenomicssystems model ==== Body Introduction The continued increase in global population, climate change and improved economic status throughout the world create a great demand for increased food productivity. Previous studies suggest that production of major crops needs to double to meet the projected demand by the year 2050, which requires an increased speed of yield improvement compared to the historical trends (Ray et al. 2013). Unfortunately, this increasing demand is in direct contrast with the decreasing rate of major crops’ yield enhancement in recent years, including rice, wheat, barley, soybean and maize (Peng et al. 2009; Zhu et al. 2010; Ray et al. 2012; Rötter et al. 2015). Finding new methods to improve major crop’s productivity has become a key challenge for plant biologists and crop breeders. With the rapid development of genomics in plant sciences, molecular breeding (i.e. marker-assisted selection and genetic engineering breeding; Rao et al. 2014) is contributing more than ever to advancements in crop improvement (Peleman and van der Voort 2003; Moose and Mumm 2008; Rao et al. 2014). In the past decades, much progress has been made by molecular breeding in developing crops with higher yield, biotic and abiotic stress resistance, and improved grain quality and plant physiology (Moose and Mumm 2008; Rao et al. 2014). However, grain yield, as a super complex trait, is determined by source, sink and flow, and their complex interaction with the environment. These components and their interactions impact crop yields in a highly non-linear manner (Chang and Zhu 2017). Modelling is the most efficient method to pinpoint the morphological and physiological parameter to improve and define the optimal values of these parameters. Although the recently proposed approach ‘breeding by rational design’, which pyramids tens of superior genes from different parents, has resulted in high-yield and superior-quality progenies (Qian et al. 2016; Zeng et al. 2017), a systematic mining and assembling of superior genes, especially those interacting genes, will likely yield even better results. Given the number of processes affecting crop yield, including canopy photosynthesis, root hydraulic and absorption property, material assimilation, metabolism, transport and utilization/storage, and the extensive interactions between these processes, what is the realistic approach to use in current breeding programmes to systematically increase crop yield? Here we propose a framework for future crop breeding, where phenomics, genomics and systems models will be the three major pillars. While extensive reviews are available in literature for each of these components (e.g. Furbank and Tester 2011; Morrell et al. 2012; Zhu et al. 2016), we discuss them as integrative components required for the future breeding of high-yielding photosynthetically efficient crops (HYPEC). Specifically, we first review the challenges of traditional crop breeding in realizing HYPEC. Then, we review the roles of plant systems models in understanding of complex biological systems and engineering for better plants, from which we argue that systems models hold great potential in guiding future crop improvement, especially with the synergy of systems models with genomics and phenomics. Finally, we propose a framework of model-guided breeding for HYPEC based on an effective combination of phenomics, genomics and systems models. Bottlenecks to breed HYPEC The stagnation in crop yield enhancement in farmers’ field in recent years suggests that the potential of traditional breeding efforts is rapidly diminishing (Long et al. 2015). Following are the major challenges facing current HYPEC breeding: Lengthy breeding cycle: In a traditional breeding programme, when a particular trait needs to be incorporated into an existing crop line, a typical practice is to find another line with the desired trait, then crossing and backcrossing it with the existing line multiple times until the trait is fixed while the existing superior traits are retained in the new line. However, this procedure is usually very time-consuming, e.g. it took 14 years to breed the line Y58S (Deng 2005), the female parental line for China’s hybrid rice cultivar Y-Liang-You 900, which reached a yield of 15.4 t ha−1 on ~7 hm2 scale in 2014 (Li et al. 2014). Difficulty of identifying breeding targets for current elite crop lines: The breeding targets here are referred to the most limiting parameters for further improvement of photosynthetic efficiency and crop yield. Traditional breeders empirically select such targets to improve based on their own breeding experience. Most of the time, they identify traits that show superior performance in other elite lines and combine them together, with the assumption that there will be additive effects of such ‘superior’ traits (Zeng et al. 2017). However, some notions about traits that were traditionally regarded as beneficial for crop yield might not be accurate. As an example, a recently theoretical study showed that decreasing, rather than increasing, leaf chlorophyll content can increase both photosynthetic light and nitrogen use efficiency (Song et al. 2017). Another example is although breeders tend to select crop lines with higher leaf area index, it was found that decreasing soybean leaf area would raise crop yields under global atmospheric change (Srinivasan et al. 2017). Difficulty of characterizing ideotypes for a defined environment: As discussed in the Introduction section, during decades of breeding, a number of ideotypes for high yields have been proposed by breeders. However, the current concepts of ideotype either differ drastically among breeders, or are physiological and architectural traits that are difficult to describe or quantify. For instance, some rice breeders consider erect panicles should be a feature of HYPEC since it can benefit canopy air circulation and enable deeper penetration of light into the canopy at noon (Chen et al. 2000), while some believe that 15–40 % higher yield can be gained by bending and maintaining panicles in a lower position inside a canopy (Setter et al. 1996). Moreover, the parameters in the current definition of ideotype are still mostly constrained to morphological traits during the grain filling stage (Peng et al. 1994; Yuan 1998; Chen et al. 2000), whereas physiological parameters and vegetative growth dynamics are rarely utilized in the breeding programmes. Furthermore, most of the parameters are related to above-ground traits, whereas the detailed root traits are rarely described. Lastly, current methods used to define ideotypes require experimental data and can therefore not be used to define ideotypes for future climate conditions, such as under elevated CO2 or temperature or altered precipitation patterns (Li et al. 2015). These issues in current ideotype definitions can be traced back to the lack of quantitative framework which can help define the contribution of different morphological or physiological parameters, to crop yield formation. Lack of knowledge regarding the quantitative genetic control of critical traits: A typical case is that although numerous theoretical studies have suggested a great advance can be made in crop yield improvement by increasing photosynthesis (Zhu et al. 2010; Gu et al. 2014), the genetic basis of leaf photosynthetic capacity under high/low light and temperature, and leaf photosynthesis response for fluctuating environmental conditions are largely unexplored and not incorporated into crop breeding (Long 2014). There are two reasons for this lack of knowledge. First, many traits are under a complex network of genetic control, including crop height, heading date, tillering and grain size (Xing and Zhang 2010; Zuo and Li 2014). Second, it is difficult to characterize desired features, including morphological traits and physiological features for both above-ground and below-ground tissues, for a large number of candidate lines throughout their growing seasons. Systems models and their roles in supporting plant science research Plant systems modelling refers to using mathematical models to quantitatively represent, integrate and simulate different physical, biochemical and physiological processes at a cell, organ, plant, population or ecosystem level. Systems model was first introduced into plant science for evaluation of canopy photosynthesis (de Wit 1965). Since then, systems models have been used to support the study of complex and highly non-linear systems, such as photosynthesis, assimilates partitioning, plant morphogenesis and soil–plant–atmosphere interaction. For example, by mathematically modelling root anatomy, 3D architecture, growth and function, the systems model OPENSIMROOT can be used to estimate the resource costs of developmental and anatomical traits, understand different root branching response patterns to soil nutrients and study among-root competition for nutrients (Postma et al. 2017). Also, systems models can guide identifying targets to engineer for desired traits in a complex system. Zhu et al. (2007) developed a kinetic model of photosynthesis in a typical C3 mesophyll cell by describing detailed metabolic processes of the Calvin Benson cycle, starch synthesis, photosynthetic carbon oxygenation pathway, triose-P export and sucrose synthesis. The optimal nitrogen partitioning pattern among enzymes was predicted given a fixed amount of total protein nitrogen, and the key enzymes which need to be overexpressed for a higher photosynthetic rate, such as Rubisco, sedoheptulose-1,7-bisphosphatase, fructose-1,6-bisphosphate aldolase and ADP-glucose pyrophosphorylase, were identified. Consistent with the prediction, overexpression of SBPase in tobacco led to increased photosynthetic rate and biomass production in the field (Rosenthal et al. 2011). Systems models were recently shown to be able to predict physiological and growth properties beyond photosynthetic metabolism. For instance, Chew et al. (2014) integrated four existing models describing different aspects of Arabidopsis growth and development to form a multi-scale framework model. By adjusting only a few parameters from the original models, the authors could quantitatively predict the metabolic, physiological and biomass dynamics during vegetative growth of different Arabidopsis accessions grown under different environmental regimes. They found that by only adjusting one model parameter, the framework model generated significantly different phenotype on leaf number, leaf size, leaf size distribution and biomass at flowering, which was further validated by transgenic plants overexpressing miR156. This successful attempt suggests that it is a reachable goal now to use systems models to design and engineer plant phenotypes from basic molecular and biochemical properties. How can systems models contribute to HYPEC breeding? After decades of research, a number of crop systems models have been developed, e.g. APSIM (McCown et al. 1996), CROPGRO (Boote et al. 1998) and DSSAT (Jones et al. 2003), and used to guide agronomical practices. Here we review how systems models contributed to crop breeding. Firstly, systems models can be used to predict crop yield. For instance, using an ensemble of 13 crop models, Li et al. (2015) predicted rice grain yields for multi-year experimental yield data at four sites with different environmental conditions with an uncertainty of <10 %. Secondly, systems models can help design cultivation practices. Recently, by using a 3D canopy photosynthesis model, Wang et al. (2017) designed a dual row planting scheme with asymmetric spacing of rows, which can decrease damage to plants and soil structure from harvest equipment with relatively little impact on canopy photosynthesis. Thirdly, systems models can guide marker-assisted selection with combination of quantitative trait loci (QTL) mapping. Gu et al. (2014) analysed the impact of genetic variation in leaf photosynthetic rate via QTL on crop biomass production using a mechanistic crop growth model GECROS in a rice introgression line population. This work highlights the potential of increasing leaf photosynthesis and crop production in rice (Long 2014). Challenges of crop systems models in guiding HYPEC breeding To date, crop systems models have not been effectively utilized to guide HYPEC breeding, which is in contrast to the success of using leaf or canopy scale photosynthesis models in guiding crop improvements, e.g. accelerating recovery from photoprotection increased tobacco biomass (Kromdijk et al. 2016) as indicated by the theoretical prediction using a canopy photosynthesis model (Zhu et al. 2004). This lack of success in using crop systems models in guiding crop breeding is mainly due to the presence of several major barriers, which are as follows. Lack of a comprehensive, multi-scale and mechanistic model describing biogenesis, function, growth and senescence of each organ in a plant growing in a community in the field: This is largely because crop yield is a complex trait. Developing a systems model covering each aspect of plant development mechanistically is far beyond the capacity of any one single research group. To tackle this challenge, a global consortium with the goal of developing such a model has been advocated (Zhu et al. 2016; Marshall-Colon et al. 2017). Difficulty in model parameterization: After model integration, the first key step of applying the framework model to a specific crop cultivar grown under a certain environment is to assign a value for each model parameter in the crop systems model. However, considering that a complete crop systems model needs to simulate crop functional and structural dynamics during a whole growth season, plant–environment interaction and even genetic and transcriptional regulatory network dynamics, its thorough parameterization is, if not impossible, laborious and expensive. This greatly limits the application of systems models in crop breeding, where thousands of candidate plants need to be evaluated to identify superior lines. Lack of information linking genomic and environmental information to model parameters: In the post-genomics era, molecular marker-based selection, genomic selection and genomic editing have become progressively faster and more efficient. Crop models can support crop molecular breeding if they can predict differences in the performance of different genotypes under different conditions (Yin et al. 2003b; Cooper et al. 2009). However, so far, the linkage between crop systems models and genomic information is still largely missing. Only a few biological processes have been modelled based on genomic information (Wilczek et al. 2009), while most of biological processes are still modelled based on macroscopic measurements, due to a lack of information on the molecular basis for these processes and the complex interaction between these processes and the environment. An emerging opportunity for achieving HYPEC with the support of crop systems models, phenomics and genomics Components needed to develop a comprehensive and mechanistic crop systems model. A mechanistic systems model of crop development and growth should be comprised of both structural and functional modules (Fig. 1). The structural module includes 3D modelling of development and growth of each organ, the assembly of a whole plant and plant community, the micro-environment in the canopy and rhizosphere. The functional module includes carbon, nutrients and water uptake and conversion in each organ, material transport in the xylem/phloem transport system and partitioning within the plant, plant–atmosphere interaction, root–soil interaction, organ senescence and material remobilization (Fig. 1A, B). To date, a number of elaborate systems models comprising the two modules have already been developed. Here we summarize the minimal set of models required to form a prototype plant-community level systems model. Firstly, modelling of organ morphogenesis must be included. Both above- and under-ground organ development can be modelled using L-system or analogues to generate and store branching structures and their relationships (Lynch et al. 1997; Fournier and Andrieu 1999; Watanabe et al. 2005; Leitner et al. 2010). Secondly, modelling of organ growth for the formed organs is required. Currently, the growth of organs is modelled based on empirical growth pattern extracted from experimental observations (Yin et al. 2003a). However, to enable models to predict growth under different environments and even future climates, mechanistic and quantitative models incorporating regulatory modules for the regulation of cell division and expansion are urgently needed and are possible now (e.g. the mechanistic modelling of cell division and growth on leaf growth; Fox et al. 2018). Thirdly, modelling of CO2 and nutrients (e.g. nitrogen, phosphate and potassium) assimilation is needed. The 3D architecture-based canopy and root models, using organ physical and physiological properties, micro-environment around plants and inside organs as input, can predict canopy photosynthesis and root absorption activities (Song et al. 2013; Xiao et al. 2016; Postma et al. 2017). Followed by photosynthetic CO2 uptake, modelling of assimilates transport, partitioning and potential remobilization (Allen et al. 2005; Yin and van Laar 2005; Thorpe et al. 2011; Chang and Zhu 2018) is also required. Finally, the available assimilates in different organs will drive the next round of organ growth and hence needs to be simulated in a complete plant-community level systems model as well. Figure 1 Elements needed for a mechanistic systems model of crop development and growth. (A) The structural (left) and functional (right) module. The structural module includes 3D modelling of each organ, the assembly of a whole plant and plant community, the micro-environment in the canopy and at the rhizosphere; the functional module includes carbon, nutrients and water uptake, assimilation and conversion in each organ, material transport within the xylem/phloem transport system and partitioning, plant–atmosphere interaction and root–soil interaction, organ senescence and material remobilization. (B) The changes of rice architecture during its life cycle. Detailed description can be found in the main text. Integration of these individual models will not be an easy task as they usually have different time steps, different types of input/output data, work on different temporal and spatial scales, and were developed using different programming languages. The key challenge now is to begin a community effort to develop a universal interface to integrate them effectively (Zhu et al. 2016). Some progress has already been made such as initial attempts of multi-scale model integration in Arabidopsis and maize (Chew et al. 2014; Postma et al. 2017). Moreover, a number of programmes have been set up with the goal of reconstructing a functioning crop plant and community of plants from the genes upward by effective model development and integration (Marshall-Colon et al. 2017; Xiao et al. 2017). High-throughput phenotyping enables morphological and physiological parameter quantification to support model parameterization and application. To guide crop breeding, crop systems models need to be well validated and further parameterized for the target crops or particular cultivars of a crop. This is a major challenge due to the difficulty of collecting comprehensive structural and functional parameters needed for crop systems models, as discussed above. For parameterization of a model simulating a simpler system, empirical relationships derived from fitting existing data can be used to convert measurable traits into model parameters. However, for an irreducible complex system, inverse problem theory needs to be applied to enable mapping from measurable traits to model parameters (Tarantola 1987). High-throughput phenotyping technology now enables researchers to collect an unprecedented quantity and quality of phenomic (i.e. morphological, physiological, biochemical and molecular trait) data for plants grown at different temporal and spatial scales and environmental conditions (Table 1). Many morphological traits previously measured manually, such as plant height and biomass, days to heading, panicle and grain features, can now be measured more rapidly and accurately with phenomics techniques (Crimmins and Crimmins 2008; Tilly et al. 2014; Yang et al. 2014). Phenomics also enables the measurement of morphological traits that are difficult to track visually and non-destructively. For example, magnetic resonance imaging (MRI) and X-ray computed tomography (CT) can be used to quantify 3D root morphological parameters in situ in soil (Metzner et al. 2015). Many physiological traits, which historically can only be measured manually, can also be measured with phenomics platforms. Yang et al. (2014) developed an automated high-throughput rice traits scorer, which can quantify grain number per panicle, grain shape, filled-grain ratio and 1000-grain weight. Dreccer et al. (2014) developed a high-throughput approach to quantify wheat stem non-structural carbohydrates (NSC) in the field using hyperspectral reflectance. Later, Wang et al. (2016) proposed a more robust approach based on infrared spectrum scanner for evaluation of stem NSC in rice. Likewise, leaf nitrogen concentration, leaf mass per area, maximum rates of RuBP carboxylation and regeneration can be measured using fresh-leaf reflectance spectroscopy (Serbin et al. 2012). High-throughput molecular level data collection is now et al. (2016) possible, as demonstrated by the high-throughput transcriptome sequencing and the emerging robotized enzymes activity measurement platforms (Gibon et al. 2004; Tombuloglu et al. 2015). A comprehensive list of these parameters that need to be collected at different developmental stages of crop growth and development and corresponding state-of-the-art phenotyping methods are summarized in Table 1. Table 1 Key parameters used in model-guided crop breeding by design and corresponding state-of-the-art phenotyping methods. Parameter Phenotyping method Throughput/accuracy Reference Environmental parameters during the whole growth season  Direct/scatter light intensity Field weather station High/high –  Air property Field weather station High/high –  Soil property Field soil sensors High/high – Whole-plant features at different growth stages  Days to heading/harvest Field repeat digital imaging High/high Crimmins and Crimmins (2008)  Canopy microclimate and photo-synthetic parameters Field canopy photosynthesis and spiration measurement system Medium/high Song et al. (2016)  3D canopy structure Stereovision Medium/high Hui et al. (2018) Parameterization from 2D images Medium/high Song et al. (2013)  Organ optical property Full-spectrum all-direction organ reflectance and transmittance meter Medium/high In development  Organ anatomy structure Nuclear magnetic resonance or microcomputed tomography Low/high Brodersen and Roddy (2016) 3D imaging Low/high Palmer et al. (2015)  Phloem/xylem (un)loading kinetics – – –  Phloem/xylem hydraulic conductance – – – Leaf features at different growth stages  Activity of photosystems Kinetic chlorophyll fluorescence im- aging systems High/high Tschiersch et al. (2017)  Stomatal conductance Thermal infrared imaging High/medium Jones (1999) Gas exchange Medium/high Juenger et al. (2005)  Nitrogen content Leaf reflectance spectroradiometer High/high Yendrek et al. (2016)  Enzymes activity of basic carbon and nitrogen metabolism Robotic enzyme activity platform High/High Gibon et al. (2004)  Hydraulic conductance Evaporative flux method, rehydration kinetics methods, etc. Medium/medium Flexas et al. (2013) Stem features during grain filling Strength In-lab prostrate tester Low/high Kashiwagi and Ishimaru (2004)  Carbohydrates storage In-lab near-infrared spectrum scanner Medium/high Wang et al. (2016) Field canopy hyperspectral reflectance High/medium Dreccer et al. (2014)  Carbohydrates metabolic activity Robotic enzyme activity platform High/high Gibon et al. (2004) Panicle features during grain filling  Grain number per panicle In-lab 2D imaging Medium/high Yang et al. (2014)  Panicle architecture In-lab 2D imaging Medium/high Crowell et al. (2014)  Photosynthesis and respiration Gas exchange Low/high In development  Endosperm cell division activity – – –  Endosperm cell energy metabolism and starch/protein synthesis activity Robotic enzyme activity platform High/high Gibon et al. (2004)  Filling pattern for grains at different panicle position Manual measurement Low/high Yang et al. (2005)  Filled-grain ratio, grain shape, grain weight In-lab machine vision-based facility High/high Yang et al. (2014) Root features at different growth stages  Growth activity – – –  Minerals absorption activity Robotic enzyme activity platform High/high Gibon et al. (2004)  Hydraulic conductance Evaporative flux method Low/high Sack and Scoffoni (2012) Rehydration kinetics methods High/medium Brodribb and Holbrook (2003)  Amount and composition of exudates High-performance liquid chromatography or gas chromatography/mass spectrometry Low/high Li et al. (2013)  3D architecture X-ray computer tomography Medium/medium Pfeifer et al. (2015) Magnetic resonance imaging Low/medium van Dusschoten et al. (2016) It is worth emphasizing here that systems models may also drive the development and expansion of phenotyping tools. Firstly, systems models can be used to identify critical morphological and physiological parameters to be screened using phenomic approaches. As an example, a model of carbon isotope discrimination and stomatal conductance suggested that leaf water use efficiency can be ‘recorded’ in the 13C/12C ratio of the leaf (Condon et al. 2002). This ratio was used to guide the development of Drysdale, a bread wheat cultivar with improved transpiration efficiency and grain yield (Condon et al. 2004). Furthermore, systems models can be used to integrate different directly measured parameters to calculate or predict new and unmeasurable parameters that represent status of growth and development, e.g. through the integration of 3D plant architecture data and leaf optical properties, canopy photosynthesis models can predict the total canopy photosynthetic CO2 uptake rate (Song et al. 2013), which is a critical physiological parameter related to HYPEC breeding. Lastly, models of plant morphology or physiology can support parameter extraction from raw phenotype data by generating large amounts of synthetic data as a training set, which was exemplified recently by using synthetic plant architecture in deep learning to enable accurate leaf counting in photographs of rosette plants (Ubbens et al. 2018). Genomics accelerates the development of crop systems models and realization of crops designed with these models. As mentioned above, a key feature of future advanced crop systems models is to predict phenotype from genotype. How can one incorporate genomic information into systems models? The first approach is to construct mapping functions between molecular markers and macroscopic model parameters. For example, heading time of individual lines in a recombination inbred population was successfully predicted under different environments based on the parameterization of four model parameters from corresponding QTL markers (Yin et al. 2005). Similarly, grain number per panicle of rice was predicted for ~1500 elite hybrid rice lines based on the number of superior alleles in each line (Huang et al. 2015). Recently, human height was accurately predicted from genotype data in a large population using a machine learning approach (Lello et al. 2018); this approach should also be applicable in the prediction of crop complex traits. Constructing a genetic regulatory network to directly predict phenotype from genotype is another approach for model parameterization. By using a gene network-based model, Wilczek et al. (2009) accurately predicted flowering time of different genotypes of Arabidopsis. Furthermore, with the genome information of Mycoplasma genitalium, a whole-cell computational model was constructed to predict a wide range of phenotypes from its genotype (Karr et al. 2012). It is foreseeable that these different approaches will be used either individually or together to obtain model parameterization with progressively improved accuracy. Genomics can accelerate identification of alleles or QTLs controlling traits identified by crops systems model as targets to produce HYPEC. Such QTLs or alleles have already been used in molecular marker-assisted breeding (Collard et al. 2005). Selection based on molecular markers outperforms traditional selection based on phenotype which can be influenced by growth environments, and has been widely used in breeding crops for biotic/abiotic resistance, high yield and superior quality, including maize (Ribaut and Ragot 2007; Prasanna et al. 2010), rice (Singh et al. 2001; Zhou et al. 2003; Jena and Mackill 2008), wheat (Gupta et al. 1999; Miedaner and Korzun 2012) and potato (Barone 2004). Genomic information is also required to support current genome editing techniques which can directly modify genomic sequences to modify particular traits or be used to perform multiple genes knock-ins, knock-downs or knock-outs simultaneously (Czarnecki et al. 2016; Zhu et al. 2017; Miao et al. 2018). Compared to marker-assisted breeding, breeding using genome editing techniques can accelerate production of HYPEC since it can potentially produce homozygous target gene(s) in one generation (Zhang et al. 2014). Finally, genomic information and the function map between model parameter and genotype will be used to guide gene or allele pyramiding (Zeng et al. 2017) or genome editing (Rodriguez-Leal et al. 2017) to realize the predicted crop ideotype. Synergy between systems models, phenomics and genomics for HYPEC breeding Here we summarize the above contents and describe the framework of systems model-guided HYPEC breeding, which will be built on the synergism of systems models, phenomics and genomics (Fig. 2). Firstly, a mapping function between model parameters and experimentally measurable parameters is needed, either by model reduction or using inverse problem theory (item 1 in Fig. 2). Secondly, high-throughput phenotyping can provide required phenomics data for parameterization (item 2.1 in Fig. 2), and genomics can help predict model parameters from genomic sequences (item 2.2 in Fig. 2). Thirdly, the parameterized model will be calibrated and/or improved by simulating crop growth under different environmental conditions (item 3 in Fig. 2). Fourthly, a well-established and parameterized systems model can be used to predict crop yield for one specific crop line grown under particular environment, and to guide field management practices, e.g. choice of crop’s growth ecological zone, decision of water and fertilizer regimes and optimization of plant spacing (item 4 in Fig. 2). Such predictions can directly benefit farmers, in terms of achieving higher output or higher ratio of output/investment. Fifthly, systems models are used to identify new breeding targets. Specifically, by parameters perturbation and in silico simulation, crop systems models can be used to identify key parameters or parameter combinations to increase yield (item 5.1 in Fig. 2). These identified critical parameters can be screened through phenomics facility assisted trait measurements (item 5.2 in Fig. 2). Sixthly, crop systems models can be used together with optimization algorithms to design optimal combinations of plant structural and functional parameters, including physiological, biochemical and molecular parameters, during the whole growth season under defined environmental conditions to gain maximal productivity (item 6.1 in Fig. 2). Then, genomic information and technology will be used to realize the designed crop ideotype (item 6.2 in Fig. 2). Undoubtedly, this framework can only be developed and improved with the participation of scientists from systems biology, molecular biology, engineering science and computational science. During this process, a close cooperation with breeders is extremely important to ensure that the model will better capture the breeder’s experience in field managements, breeding targets identification or ideotype definition. Figure 2 A conceptual framework of model-guided breeding for HYPEC with an effective combination of systems models, genomics and phenomics. 1, mapping between model parameters and experimentally measurable parameters; 2.1, phenomics facilitated quantification of measurable parameters; 2.2, direct prediction of model parameters from genome; 3, prediction of crop growth using a parameterized systems model; 4, prediction of crop yield and selection of optimal field management practice using mechanistic systems models; 5.1, systematic identification of breeding targets; 5.2, phenomics assisted screening of breeding targets; 6.1, defining of crop ideotype; 6.2, genomics assisted realization of crop ideotype. Conclusion Improving photosynthetic efficiency over a whole growing season is now recognized as a major option to drastically improve crop yield, for which all the sink and flow related processes also need to be simultaneously considered to achieve the expected gain in crop yield potential. In this aspect, many breeding organizations, such as CIMMYT, have recognized this and created physiology-based breeding programmes to simultaneously improve source-, sink- and flow-related traits (Reynolds and Langridge 2016). This work presents the rationale and a strategy to expedite HYPEC breeding by combining systems models, phenomics and genomics techniques in the same breeding programme. We envisage that effective combination of these different techniques, which are, at this point, mostly scattered in different research or breeding organizations, will create enormous synergy and help catalyse another round of rapid gain in crop yield potential in the coming decades. Acknowledgements The authors thank Y. Wang, H. Zhao and Z. Zhou for assistance in figure preparation. Sources of Funding This research was financially supported by the Strategic Priority Research Programme of the Chinese Academy of Sciences (XDB27020105; XDA08020301), Bill and Melinda Gates Foundation (OPP1172157) and the National High Technology Development Programme (2014AA101601). Contribution by authors X.G.Z. planned the paper, T.G.C., S.C., Q.F.S., S.P. and X.G.Z. wrote the paper. Conflict of Interest None declared. ==== Refs Literature Cited Allen MT , Prusinkiewicz P , DeJong TM 2005 Using L-systems for modeling source-sink interactions, architecture and physiology of growing trees: the L-PEACH model . New Phytologist 166 :869 –880 . Barone A 2004 Molecular marker-assisted selection for potato breeding . American Journal of Potato Research 81 :111 –117 . Boote KJ , Jones JW , Hoogenboom G 1998 Simulation of crop growth: CROPGRO - soybean model In: Peart RM , Curry RB , eds. Agricultural systems modeling & simulation . New York : Marcel Dekker , 1 –42 . Brodersen CR , Roddy AB 2016 New frontiers in the three-dimensional visualization of plant structure and function . American Journal of Botany 103 :184 –188 .26865119 Brodribb TJ , Holbrook NM 2003 Stomatal closure during leaf dehydration, correlation with other leaf physiological traits . Plant Physiology 132 :2166 –2173 .12913171 Chang TG , Zhu XG 2017 Source-sink interaction: a century old concept under the light of modern molecular systems biology . Journal of Experimental Botany 68 :4417 –4431 .28338782 Chang T , Zhu X 2018 Systematic optimization of whole plant carbon nitrogen interaction (WACNI) to support crop design for greater yield . bioRxiv . doi: 10.1101/286112 . Chen W , Xu Z , Zhang W , Zhang L , Yang S 2000 Creation of new plant type and breeding rice for super high yield . Acta Agronomica Sinica 27 :665 –672 . Chew YH , Wenden B , Flis A , Mengin V , Taylor J , Davey CL , Tindal C , Thomas H , Ougham HJ , De Reffye P 2014 Multiscale digital Arabidopsis predicts individual organ and whole-organism growth . Proceedings of the National Academy of Sciences of the United States of America 111 :E4127 –E4136 .25197087 Collard B , Jahufer M , Brouwer J , Pang E 2005 An introduction to markers, quantitative trait loci (QTL) mapping and markerassisted selection for crop improvement: the basic concepts . Euphytica 142 :169 –196 . Condon AG , Richards R , Rebetzke G , Farquhar G 2002 Improving intrinsic water-use efficiency and crop yield . Crop Science 42 :122 –131 .11756262 Condon AG , Richards R , Rebetzke G , Farquhar G 2004 Breeding for high water-use efficiency . Journal of Experimental Botany 55 :2447 –2460 .15475373 Cooper ME , Fred A , Hammer GL , Podlich DW , Messina C 2009 Modeling QTL for complex traits: detection and context for plant breeding . Current Opinion in Plant Biology 12 :231 –240 .19282235 Crimmins MA , Crimmins TM 2008 Monitoring plant phenology using digital repeat photography . Environmental Management 41 :949 –958 .18288519 Crowell S , Falcão AX , Shah A , Wilson Z , Greenberg AJ , McCouch SR 2014 High-resolution inflorescence phenotyping using a novel image-analysis pipeline, PANorama . Plant Physiology 165 :479 –495 .24696519 Czarnecki O , Bryan AC , Jawdy SS , Yang X , Cheng ZM , Chen JG , Tuskan GA 2016 Simultaneous knockdown of six non-family genes using a single synthetic RNAi fragment in Arabidopsis thaliana . Plant Methods 12 :16 .26889205 Deng QY 2005 Breeding of the PTGMS Line Y58S with wide adaptability in rice . Hybrid Rice 20 :15 –18 . de Wit CT 1965 Photosynthesis of leaf canopies . Centre for Agricultural Publications and Documentation , Wagenirtgen . Dreccer MF , Barnes LR , Meder R 2014 Quantitative dynamics of stem water soluble carbohydrates in wheat can be monitored in the field using hyperspectral reflectance . Field Crops Research 159 :70 –80 . Flexas J , Scoffoni C , Gago J , Sack L 2013 Leaf mesophyll conductance and leaf hydraulic conductance: an introduction to their measurement and coordination . Journal of Experimental Botany 64 :3965 –3981 .24123453 Fournier C , Andrieu B 1999 ADEL-maize: an L-system based model for the integration of growth processes from the organ to the canopy. Application to regulation of morphogenesis by light availability . Agronomie 19 :313 –327 . Fox S , Southam P , Pantin F , Kennaway R , Robinson S , Castorina G , Sánchez-Corrales YE , Sablowski R , Chan J , Grieneisen V , Marée AFM , Bangham JA , Coen E 2018 Spatiotemporal coordination of cell division and growth during organ morphogenesis . PLoS Biology 16 :e2005952 .30383040 Furbank RT , Tester M 2011 Phenomics--technologies to relieve the phenotyping bottleneck . Trends in Plant Science 16 :635 –644 .22074787 Gibon Y , Blaesing OE , Hannemann J , Carillo P , Höhne M , Hendriks JH , Palacios N , Cross J , Selbig J , Stitt M 2004 A robot-based platform to measure multiple enzyme activities in Arabidopsis using a set of cycling assays: comparison of changes of enzyme activities and transcript levels during diurnal cycles and in prolonged darkness . The Plant Cell 16 :3304 –3325 .15548738 Gu J , Yin X , Stomph T-J , Struik PC 2014 Can exploiting natural genetic variation in leaf photosynthesis contribute to increasing rice productivity? A simulation analysis . Plant, Cell & Environment 37 :22 –34 . Gupta PK , Varshney RK , Sharma PC , Ramesh B 1999 Molecular markers and their applications in wheat breeding . Plant Breeding 118 :369 –390 . Huang X , Yang S , Gong J , Zhao Y , Feng Q , Gong H , Li W , Zhan Q , Cheng B , Xia J 2015 Genomic analysis of hybrid rice varieties reveals numerous superior alleles that contribute to heterosis . Nature Communications 6 :6258 . Hui F , Zhu J , Hu P , Meng L , Zhu B , Guo Y , Li B , Ma Y 2018 Imagebased dynamic quantification and high-accuracy 3D evaluation of canopy structure of plant populations . Annals of Botany 121 :1079 –1088 .29509841 Jena K , Mackill D 2008 Molecular markers and their use in markerassisted selection in rice . Crop Science 48 :1266 –1276 . Jones H 1999 Use of thermography for quantitative studies of spatial and temporal variation of stomatal conductance over leaf surfaces . Plant, Cell & Environment 22 :1043 –1055 . Jones JW , Hoogenboom G , Porter C , Boote K , Batchelor W , Hunt L , Wilkens P , Singh U , Gijsman A , Ritchie J 2003 The DSSAT cropping system model . European Journal of Agronomy 18 :235 –265 . Juenger TE , Mckay JK , Hausmann N , Keurentjes JJ , Sen S , Stowe KA , Dawson TE , Simms EL , Richards JH 2005 Identification and characterization of QTL underlying whole-plant physiology in Arabidopsis thaliana: δ13C, stomatal conductance and transpiration efficiency . Plant, Cell & Environment 28 :697 –708 . Karr JR , Sanghvi JC , Macklin DN , Gutschow MV , Jacobs JM , Bolival B , Assadgarcia N , Glass JI , Covert MW 2012 A whole-cell computational model predicts phenotype from genotype . Cell 150 :389 –401 .22817898 Kashiwagi T , Ishimaru K 2004 Identification and functional analysis of a locus for improvement of lodging resistance in rice . Plant Physiology 134 :676 –683 .14739343 Kromdijk J , Głowacka K , Leonelli L , Gabilly ST , Iwai M , Niyogi KK , Long SP 2016 Improving photosynthesis and crop productivity by accelerating recovery from photoprotection . Science 354 :857 –861 .27856901 Leitner D , Klepsch S , Bodner G , Schnepf A 2010 A dynamic root system growth model based on L-Systems . Plant and Soil 332 :177 –192 . Lello L , Avery SG , Tellier L , Vazquez AI , de los Campos G , Hsu SDH 2018 Accurate genomic prediction of human height . Genetics 210 :477 –497 .30150289 Li T , Hasegawa T , Yin X , Zhu Y , Boote K , Adam M , Bregaglio S , Buis S , Confalonieri R , Fumoto T 2015 Uncertainties in predicting rice yield by current crop models under a wide range of climatic conditions . Global Change Biology 21 :1328 –1341 .25294087 Li X-G , Zhang T-L , Wang X-X , Hua K , Zhao L , Han Z-M 2013 The composition of root exudates from two different resistant peanut cultivars and their effects on the growth of soil-borne pathogen . International Journal of Biological Sciences 9 :164 .23412138 Li J , Zhang Y , Wu J , Lin S , Zhou P , Deng Q 2014 High-yielding cultural techniques of super hybrid rice YLY 900 yielded 15.40 t/hm2 on a 6.84 hm2 scale . China Rice 20 :1 –4 . Long SP 2014 We need winners in the race to increase photosynthesis in rice, whether from conventional breeding, biotechnology or both . Plant Cell and Environment 37 :19 –21 . Long SP , Marshall-Colon A , Zhu X-G 2015 Meeting the global food demand of the future by engineering crop photosynthesis and yield potential . Cell 161 :56 –66 .25815985 Lynch JP , Nielsen KL , Davis RD , Jablokow AG 1997 SimRoot: modelling and visualization of root systems . Plant and Soil 188 :139 –151 . Marshall-Colon A , Long SP , Allen DK ,Allen G , Beard DA , Benes B , Von Caemmerer S , Christensen AJ , Cox DJ , Hart JC 2017 Crops in silico: generating virtual crops using an integrative and multi-scale modeling platform . Frontiers in Plant Science 8 :786 .28555150 McCown R , Hammer G , Hargreaves J , Holzworth D , Freebairn D 1996 APSIM: a novel software system for model development, model testing and simulation in agricultural systems research . Agricultural Systems 50 :255 –271 . Metzner R , Eggert A , van Dusschoten D , Pflugfelder D , Gerth S , Schurr U , Uhlmann N , Jahnke S 2015 Direct comparison of MRI and X-ray CT technologies for 3D imaging of root systems in soil: potential and challenges for root trait quantification . Plant Methods 11 :17 .25774207 Miao C , Xiao L , Hua K , Zou C , Zhao Y , Bressan AR , Zhu J-K 2018 Mutations in a subfamily of abscisic acid receptor genes promote rice growth and productivity . Proceedings of the National Academy of Sciences of the United States of America 115 :6058 –6063 .29784797 Miedaner T , Korzun V 2012 Marker-assisted selection for disease resistance in wheat and barley breeding . Phytopathology 102 :560 .22568813 Moose SP , Mumm RH 2008 Molecular plant breeding as the foundation for 21st century crop improvement . Plant Physiology 147 :969 –977 .18612074 Morrell PL , Buckler ES , Ross-Ibarra J 2012 Crop genomics: advances and applications . Nature Reviews Genetics 13 :85 –96 . Palmer WM , Martin AP , Flynn JR , Reed SL , White RG , Furbank RT , Grof CP 2015 PEA-CLARITY: 3D molecular imaging of whole plant organs . Scientific Reports 5 . Peleman JD , van der Voort JR 2003 Breeding by design . Trends in Plant Science 8 :330 –334 .12878017 Peng S , Khush G , Cassman K 1994 Evolution of the new plant ideotype for increased yield potential . In: Cassman K.G ed. Breaking the yield barrier. Proceedings of a workshop on rice yield potential in favourable environments . Philippines : International Rice Research Institute , 5 –20 . Peng S , Tang Q , Zou Y 2009 Current status and challenges of rice production in China . Plant Production Science 12 :3 –8 . Pfeifer J , Kirchgessner N , Colombi T , Walter A 2015 Rapid phenotyping of crop root systems in undisturbed field soils using X-ray computed tomography . Plant Methods 11 :41 .26322118 Postma JA , Kuppe C , Owen MR , Mellor N , Griffiths M , Bennett MJ , Lynch JP , Watt M 2017 OpenSimRoot: widening the scope and application of root architectural models . New Phytologist 215 :1274 –1286 . Prasanna BM , Pixley K , Warburton ML , Xie CX 2010 Molecular marker-assisted breeding options for maize improvement in Asia . Molecular Breeding 26 :339 –356 . Qian Q , Guo LB , Smith SM , Li JY 2016 Breeding high-yield superior quality hybrid super rice by rational design . National Science Review 3 :283 –294 . Rao Y , Li Y , Qian Q 2014 Recent progress on molecular breeding of rice in China . Plant Cell Reports 33 :551 –564 .24442397 Ray DK , Mueller ND , West PC , Foley JA 2013 Yield trends are insufficient to double global crop production by 2050 . PLoS One 8 :e66428 .23840465 Ray DK , Ramankutty N , Mueller ND , West PC , Foley JA 2012 Recent patterns of crop yield growth and stagnation . Nature Communications 3 :1293 . Reynolds M , Langridge P 2016 Physiological breeding . Current Opinion in Plant Biology 31 :162 –171 .27161822 Ribaut JM , Ragot M 2007 Marker-assisted selection to improve drought adaptation in maize: the backcross approach, perspectives, limitations, and alternatives . Journal of Experimental Botany 58 :351 –360 .17158111 Rodriguez-Leal D , Lemmon ZH , Man J , Bartlett ME , Lippman ZB 2017 Engineering quantitative trait variation for crop improvement by genome editing . Cell 171 :470 –480 .e8 .28919077 Rosenthal DM , Locke AM , Khozaei M , Raines CA , Long SP , Ort DR 2011 Over-expressing the C3 photosynthesis cycle enzyme sedoheptulose-1-7 bisphosphatase improves photosynthetic carbon gain and yield under fully open air CO2 fumigation (FACE ). BMC Plant Biology 11 :123 .21884586 Rötter RP , Tao F , Höhn JG , Palosuo T 2015 Use of crop simulation modelling to aid ideotype design of future cereal cultivars . Journal of Experimental Botany 66 :3463 –3476 .25795739 Sack L , Scoffoni C 2012 Measurement of leaf hydraulic conductance and stomatal conductance and their responses to irradiance and dehydration using the evaporative flux method (EFM ). Journal of Visualized Experiments 70 :4179 . Serbin SP , Dillaway DN , Kruger EL , Townsend PA 2012 Leaf optical properties reflect variation in photosynthetic metabolism and its sensitivity to temperature . Journal of Experimental Botany 63 :489 –502 .21984647 Setter TL , Conocono EA , Egdane JA 1996 Possibility of increasing yield potential of rice by reducing panicle height in the canopy. II. Canopy photosynthesis and yield of isogenic lines . Functional Plant Biology 23 :161 –169 . Singh S , Sidhu J , Huang N , Vikal Y , Li Z , Brar D , Dhaliwal H , Khush G 2001 Pyramiding three bacterial blight resistance genes (xa5, xa13 and Xa21) using marker-assisted selection into indica rice cultivar PR106 . Theoretical and Applied Genetics 102 :1011 –1015 . Song Q , Wang Y , Qu M , Ort DR , Zhu X 2017 The impact of modifying photosystem antenna size on canopy photosynthetic efficiency . Plant, Cell & Environment 40 :2946 –2957 . Song Q , Xiao H , Xiao X , Zhu X-G 2016 A new canopy photosynthesis and transpiration measurement system (CAPTS) for canopy gas exchange research . Agricultural and Forest Meteorology 217 :101 –107 . Song QF , Zhang GL , Zhu XG 2013 Optimal crop canopy architecture to maximise canopy photosynthetic CO2 uptake under elevated CO2 - a theoretical study using a mechanistic model of canopy photosynthesis . Functional Plant Biology 40 :109 –124 . Srinivasan V , Kumar P , Long SP 2017 Decreasing, not increasing, leaf area will raise crop yields under global atmospheric change . Global Change Biology 23 :1626 –1635 .27860122 Tarantola A 1987 Inverse problem theory. Methods for data fitting and model parameter estimation . Physics of the Earth & Planetary Interiors 57 :350 –351 . Thorpe MR , Lacointe A , Minchin PE 2011 Modelling phloem transport within a pruned dwarf bean: a 2-source-3-sink system . Functional Plant Biology 38 :127 –138 .32480869 Tilly N , Hoffmeister D , Cao Q , Huang S , Lenz-Wiedemann V , Miao Y , Bareth G 2014 Multitemporal crop surface models: accurate plant height measurement and biomass estimation with terrestrial laser scanning in paddy rice . Journal of Applied Remote Sensing 8 :083671 . Tombuloglu G , Tombuloglu H , Sakcali MS , Unver T 2015 Highthroughput transcriptome analysis of barley (Hordeum vulgare) exposed to excessive boron . Gene 557 :71 –81 .25498907 Tschiersch H , Junker A , Meyer RC , Altmann T 2017 Establishment of integrated protocols for automated high throughput kinetic chlorophyll fluorescence analyses . Plant Methods 13 :54 .28690669 Ubbens J , Cieslak M , Prusinkiewicz P , Stavness I 2018 The use of plant models in deep learning: an application to leaf counting in rosette plants . Plant Methods 14 :6 .29375647 van Dusschoten D , Metzner R , Kochs J , Postma JA , Pflugfelder D , Bühler J , Schurr U , Jahnke S 2016 Quantitative 3D analysis of plant roots growing in soil using magnetic resonance imaging . Plant Physiology 170 :1176 –1188 .26729797 Wang Y , Song QF , Jaiswal D , de Souza AP , Long SP , Zhu XG 2017 Development of a three-dimensional ray-tracing model of sugarcane canopy photosynthesis and its application in assessing impacts of varied row spacing . Bioenergy Research 10 :626 –634 . Wang DR , Wolfrum EJ , Virk P , Ismail A , Greenberg AJ , McCouch SR 2016 Robust phenotyping strategies for evaluation of stem non-structural carbohydrates (NSC) in rice . Journal of Experimental Botany 67 :6125 –6138 .27707775 Watanabe T , Hanan JS , Room PM , Hasegawa T , Nakagawa H , Takahashi W 2005 Rice morphogenesis and plant architecture: measurement, specification and the reconstruction of structural development by 3D architectural modelling . Annals of Botany 95 :1131 –1143 .15820987 Wilczek AM , Roe JL , Knapp MC , Cooper MD , Lopezgallego C , Martin LJ , Muir CD , Sim S , Walker A , Anderson J 2009 Effects of genetic perturbation on seasonal life history plasticity . Science 323 :930 –934 .19150810 Xiao Y , Tholen D , Zhu X-G 2016 The influence of leaf anatomy on the internal light environment and photosynthetic electron transport rate: exploration with a new leaf ray tracing model . Journal of Experimental Botany 67 :6021–6035 . Xiao Y , Chang T , Song Q , Wang S , Tholen D , Wang Y , Xin C , Zheng G , Zhao H , Zhu XG 2017 ePlant for quantitative and predictive plant science research in the big data era . Quantitative Biology 5 :260 –271 . Xing Y , Zhang Q 2010 Genetic and molecular bases of rice yield . Annual Review of Plant Biology 61 :421 –442 . Yang W , Guo Z , Huang C , Duan L , Chen G , Jiang N , Fang W , Feng H , Xie W , Lian X 2014 Combining high-throughput phenotyping and genome-wide association studies to reveal natural genetic variation in rice . Nature Communications 5 :5087 . doi: 10.1038/ncomms6087 . Yang J , Zhang J , Wang Z , Liu K , Wang P 2005 Post-anthesis development of inferior and superior spikelets in rice in relation to abscisic acid and ethylene . Journal of Experimental Botany 57 :149 –160 .16330527 Yendrek C , Tomaz T , Montes CM , Cao Y , Morse AM , Brown PJ , McIntyre L , Leakey A , Ainsworth E 2016 High-throughput phenotyping of maize leaf physiology and biochemistry using hyperspectral reflectance . Plant Physiology 173 :614 –626 .28049858 Yin X , Goudriaan J , Lantinga EA , Vos J , Spiertz HJ 2003a A flexible sigmoid function of determinate growth . Annals of Botany 91 :361 –371 .12547689 Yin X , Stam P , Kropff MJ , Schapendonk AH 2003b Crop modeling, QTL mapping, and their complementary role in plant breeding . Agronomy Journal 95 :90 –98 . Yin X , Struik PC , van Eeuwijk FA , Stam P , Tang J 2005 QTL analysis and QTL-based prediction of flowering phenology in recombinant inbred lines of barley . Journal of Experimental Botany 56 :967 –976 .15710636 Yin X , van Laar H 2005 Crop systems dynamics: an ecophysiological model of genotype-by-environment interactions (GECROS ). Wageningen, The Netherlands : Wageningen Academic Publishers . Yuan L 1998 Hybrid rice breeding for super high yield . In: Denning GL , Mew TW , eds. China and IRRI: Improving China’s rice productivity in the 21st century Manila (Philippines ): International Rice Research Institute 10 –12 . Zeng DL , Tian ZX , Rao YC , Dong GJ , Yang YL , Huang LC , Leng YJ , Xu J , Sun C , Zhang GH , Hu J , Zhu L , Gao ZY , Hu XM , Guo LB , Xiong GS , Wang YH , Li JY , Qian Q 2017 Rational design of high-yield and superior-quality rice . Nature Plants 3 :17031 . doi: 10.1038/nplants.2017.31 .28319055 Zhang H , Zhang J , Wei P , Zhang B , Gou F , Feng Z , Mao Y , Yang L , Zhang H , Xu N 2014 The CRISPR/Cas9 system produces specific and homozygous targeted gene editing in rice in one generation . Plant Biotechnology Journal 12 :797 –807 .24854982 Zhou P , Tan Y , He Y , Xu C , Zhang Q 2003 Simultaneous improvement for four quality traits of Zhenshan 97, an elite parent of hybrid rice, by molecular marker-assisted selection . Theoretical and Applied Genetics 106 :326 –331 .12582859 Zhu XG , de Sturler E , Long SP 2007 Optimizing the distribution of resources between enzymes of carbon metabolism can dramatically increase photosynthetic rate: a numerical simulation using an evolutionary algorithm . Plant Physiology 145 :513 –526 .17720759 Zhu X-G , Long SP , Ort DR 2010 Improving photosynthetic efficiency for greater yield . Annual Review of Plant Biology 61 :235 –261 . Zhu XG , Lynch JP , LeBauer DS , Millar AJ , Stitt M , Long SP 2016 Plants in silico: why, why now and what?—an integrative platform for plant systems biology research . Plant, Cell & Environment 39 :1049 –1057 . Zhu XG , Ort DR , Whitmarsh J , Long SP 2004 The slow reversibility of photosystem II thermal energy dissipation on transfer from high to low light may cause large losses in carbon gain by crop canopies: a theoretical analysis . Journal of Experimental Botany 55 :1167 –1175 .15133059 Zhu Q , Yu S , Zeng D , Liu H , Wang H , Yang Z , Xie X , Shen R , Tan J , Li H 2017 Development of “purple endosperm rice” by engineering anthocyanin biosynthesis in the endosperm with a high-efficiency transgene stacking system . Molecular Plant 10 :918 –929 .28666688 Zuo J , Li J 2014 Molecular dissection of complex agronomic traits of rice: a team effort by Chinese scientists in recent years . National Science Review 1 :253 –276 .