Welcome to Chinese Agricultural Science Bulletin,

Chinese Agricultural Science Bulletin ›› 2010, Vol. 26 ›› Issue (5): 55-62.

• 23 • Previous Articles     Next Articles

Probe into Analysis Method of Plant Trait Series Data Models

  

  • Received:2009-10-16 Revised:2009-11-29 Online:2010-03-05 Published:2010-03-05

Abstract:

This paper defines the plant trait concept from a new angle, i.e. the basic trait features of heredity, course feature, fluctuation, learning character and divergence, discusses on the basis of these features the mathematical and statistical concepts and methods, e.g. the time series analysis, that might be introduced into the trait studies, and demonstrates the feasibility of this thinking and method through the studies with specimen on the traits such as quality and output of the proper xenia rape, autogamic rice and hybrid rice. The study on the substance accumulating courses of the rapeseed demonstrates that the accumulation of all the other substances except for the oil is a nonlinear course and obviously is the cubic function of the independent variables. The simple and function smoothing models of the soluble carbohydrate (SC), OIL and dry weight (DW) series are extremely significant, so are the linear models of the protein (PRO) and cellulose. However, the adenosine triphosphate (ATP), the basic energy supplier for the metabolism of many substances, is not significant, indicating that the process involved is extremely complicated and no suitable quantitative model can be obtained at present. The data analysis of the rice leaf area growth by means of the time series analysis software demonstrates that the result is apparently different from that of the rapeseed data. The rapeseed falls either in the exponential smoothing model or the linear trend model. Yet, of the 15 rice series, only 6 demonstrate the exponential smoothing model or linear trend model, and the rest exponential smoothing model with damping trend. Except for the sterile line that is of exceptional features, the restoring line, the filial generation I line, the multi-spike line and the big spike line in terms of single leaf weight all fall in the exponential smoothing model with damping trend. This result demonstrates the basic rule of changes in the rice leaf weight. Besides, the restoring line model and the filial generation (F1) line model are identical in three traits. In the simulation of the hybrid rice series with the restoring line and the sterile line, the corresponding variables of the restoring line play the major role, the sterile line has a secondary function only, demonstrating that the restoring line exerts a decisive function on the hybrid rice so far as the growth features of the rice leaf is concerned.