Genetic analysis model of forest based on space and competition effects

Objective  To improve the accuracy in genetic analysis of forest establishing an analysis model based on space and competition effects.

Method  The data was simulated by R software and its package breedR. The additive effect and neighbor competition effect were fitted using XFA1 structure and the spatial effect was fitted using AR1 structure for both simulated and measured data. Four models (randomized block design model, RCBM; spatial model, SM; spatial with measured error model, SUM; spatial and competition model, SCM) were established and analyzed using ASReml to estimate genetic parameters for comparison.

Result  The estimated results showed that SCM was the best model for the simulation data. SCM greatly reduced the random error variance from 7.56 (RCBM) and 5.72 (SUM) to 3.13 (SCM), decreased by 58.6% and 45.3%, respectively. SCM could estimate the genetic variance of competition from surrounding neighbors. The individual heritability assessed by SCM was around 0.40, higher than those of RCBM (0.24) and SUM (0.30). SCM obtained stable estimated results under different settings of initial values for parameters. Furthermore, for the measured data, the estimated results were consistent with the simulation data.

Conclusion  SCM is a new individual-tree mixed model and could be used for genetic analysis of forest.