Effects Of Stalk Tissue And Lignin Metabolism On The Lodging Resistance Of Rice

In order to realize the goal of breeding the fourth super rice varieties whose yield break 1000kg per mu, academician Yuan Longping put forward the ideas that breeding long-stalked rice varieties to increase the biological yield. However, plant height and lodging-resistance is the prominent contradiction. Therefore, breeding long-stalked and lodging-resistance rice varieties, is one of the key steps of realizing the fourth super rice goal. Many factors affect the lodging, many researchers have established the different evaluation methods of the plant type from different ways. So, we need a simple and accurate evaluation method to predict the ideal plant type of rice, especially under the background of the breeding of long-stalked and lodging-resistance super rice varieties.The cell wall is the unique structure of plant, which is essential bearing structure for crop stalks. In the process of cell wall lignifications, the synthesis of lignin in the cell wall has a great influence on the lodging strength and the lodging index. During the developmental stages of rice, cell wall lignifications among the varieties show different changing tendency, which lead to a series of changes of micro structure, and make a great influence on the bending strength of rice stem. In the metabolism of lignin, aspartate aminotransferase (PAL), 4-coumaric acid:CoA-ligase (4-CL), cinnamyl-alcohol dehydrogenase (CAD), polyphenol oxidase (PPO) are key enzymes, whose activity effects the synthesis of lignin. The relation between the active enzyme and lignin content is nonlinear, so it needs a new method to analyze. Here, we provide a new model to illuminate the feature on breeding of rice varieties and the bearing mechanism.There are many mathematical models to analyze plant lodging resistance. Such as correlation analysis and multiple regression analysis, whose defection in prediction and nonlinear fitting impends the application. So we adopt neural network for its powerful function for pattern recognition and strong nonlinear fitting ability.This subject adopts plant anatomy, histochemistry and cell chemistry and biochemistry methods to research stem bending force and lodging index of 15 different rice varieties in its development period. Then study the atomic result through the correlation analysis, multiple regression analysis (MLR) and artificial neural network (RBF) analysis. At last set up three kinds of evaluation mechanism to evaluate the stem structure, by which we identify the optimal predictive model and the best plant type. The comparision of the evaluate model shows that the residual of RBF neural network and the average relative error is minimum, is 0.04 and 7.07% respectively. The residual average value and the average relative error of MLR was significantly higher than that of RBF network, it’s prediction is low in precision, and the average value of the residual error and relative error are 0.11 and 11.27%, about 2 times as much as the RBF neural network prediction. Then by RBF neural network prediction, determine the best rice plant type:plant height 115cm-125cm, cortical thickness 0.428, the number of vascular bundle 13.3,vascular bundle proportion 77%, inverted three internodes length 16.5cm; plant height is 130cm-140cm, cortical thickness is 0.368cm, the number of vascular bundle is 13.0, vascular bundle proportion is 73.85%, inverted three internodes length 17.6cm;plant height 150cm-160cm, cortical thickness 0.494cm, the number of vascular bundle 12.5, vascular bundle proportion76.65%, inverted three internodes length 18.3cm; plant height 160cm-180cm, cortical thickness 0.62cm, the number of vascular bundle 13.4, vascular bundle proportion75.95%, inverted three internodes length 17.6cm.Measuring the cell wall lignin content and its related enzyme activity, analyzing the content change of lignin and the correspond change of bearing strength and lodging index, we found that lignification affect the material transport and the change of center of gravity, which lead to great bearing strength variation and small lodging index variation in the later grain filling stage. Using RBF neural network’s strong nonlinear fitting ability and the radial basis function centers determined by fuzzy k-means clustering method, we analyze the relationship lignin content and its related enzyme activity which determine the influence factors of lignin metabolism among different materials in different periods. The result shows that rice varieties with high bearing strength and small lodging index have higher lignin content and the activities of PAL, TAL, CAD, and 4CL than others. The lignin content was negatively correlated with the actual lodging ratio (r=-0.914, P<0.01). According to RBF neural network function center and variance, we can conclude that activity of 4-CL is higher than others, the PPO and PAL are close, lignin synthesis is at the preparation stage in the heading period. In the early grain filling stage, PAL activity is higher, and other enzyme activity began to increase and strength between enzyme activity, lignin synthesis step into the synthetic period. In the middle grain filling stage the activity of the enzyme is highest and the stem is strongest. At the later grain filling stage, the enzyme activity began to decline except the PPO, during which PAL decrease is most apparent and 4CL still maintain at a certain level, the lignification progress still go on.This paper build the rice stem morphological characteristics and rice lodging and stalk lignin metabolism model, aim to establish a fast and accurate plant lodging evaluation mechanism. Through the analysis of straw lignin metabolism to explore the microscopic mechanism of stem lodging, we could provide a theory basis for the breeding of long-stalked and lodging-resistance super rice varieties.