同源四倍体及其原二倍体黑皮冬瓜光合光响应模型筛选

为了筛选出最适合黑皮冬瓜Benincasa hispida (Thunb.) Cogn.的光合光响应模型,为其育种提供理论依据,以同源四倍体及其原二倍体黑皮冬瓜为试验材料,对8种经典的光合光响应模型适用性进行了比较分析。结果表明：二次多项式能够表现出光抑制情况,但在拟合过程中出现暗呼吸速率为正值、光补偿点为负值及无法解释当光强达到饱和后光合速率快速下降的问题;直角双曲线、非直角双曲线及指数函数Ⅰ、指数函数Ⅱ无法直接求取光饱和点、光补偿点,结合常用的光饱和点的计算方法得到的光饱和点与实测值均存在较大的偏差,且指数函数Ⅱ在计算光饱和点时表现出明显的人为性,也无法拟合光抑制情况,但4种模型拟合得到的光补偿点均与实测值相差不大;指数修正模型因系数β为负值,无法求取四倍体黑皮冬瓜材料的光饱和点和最大净光合速率,且拟合得到的四倍体黑皮冬瓜的光补偿点明显低于实测值;直角双曲线修正模型计算得到的暗呼吸速率及二倍体黑皮冬瓜的光饱和点明显低于实测值,但获得的四倍体及其二倍体的最大净光合速率与实测值最接近,说明其在拟合最大净光合速率上有优势;整体上分段函数计算得到的黑皮冬瓜的各光合参数与实测值最为接近,与实测值的平均相对误差最小,也能很好的拟合发生光抑制部分的光响应曲线。分段函数拟合同源四倍体及其原二倍体黑皮冬瓜光合光响应曲线效果较其他模型效果好,分段函数模型为黑皮冬瓜最适合的光合光响应模型。

Fitting Models of Photosynthetic and Light-response for Autotetraploid and Diploid of Benincasa hispida (Thunb.) Cogn.

The purpose of the experiment was to choose fitting models of photosynthetic and light response for autotetraploid and diploid Benincasa hispida (Thunb.) Cogn., and to provide a theoretical basis for its breeding. Eight extensively applied models of photosynthetic and light response were compared to choose the most suitable light response curve model for autotetraploid and diploid B. hispida. The binomial regression model could deal with the photoinhibiton phenomenon, but it existed some problems during fitting, such as the respiration rate (R_day) value was above zero, the light compensation point (LCP) value was below zero, and impossible to explain the rapid decline of photosynthetic rate when light intensity reached saturation. The model of rectangular hyperbola, nonrectangular hyperbola, exponential functionⅠand exponential functionⅡcould not calculate LSP and LCP value directly, while the value calculated by combining other methods existed large deviation compared with the measured value, the exponential function Ⅱ showed obvious artificiality when calculating LSP value, and could not explain photoinhibition phenomenon. However, the values of LCP calculated by the four models were no large deviation with the measured values. Because of the coefficient β of the modified exponential function model was negative, the LSP and P_nmax value of the autotetraploid B. hispida cannot be calculated, and the LCP value of autotetraploid B. hispida was also obviously lower than the measured value. The R_day and LSP value of diploid B. hispida fitted by the modified rectangular hyperbola model were obviously lower than measured values, but the P_nmax value of autotetraploid and diploid B. hispida was the closest to the measured value, and it had advantage in fitting the P_nmax by the modified rectangular hyperbola model. The average relative error (RE) of all photosynthetic parameters and measured values was lowest fitted by the subsection function, and was also the closest to the measured value, and could deal with the photoinhibiton phenomenon well. Subsection function had better a fitting effect than other models, and the most suitable photosynthetic model of B. hispida was subsection function.