Spatial analysis increases efficiency of progeny testing of Chinese fir

Abstract

We used spatial, global trend and post-blocking analysis to examine the effectiveness of a progeny trial in a tree breeding program for Chinese fir (Cunninghamia lanceolata (Lamb.) Hook) on a hilly site with an environmental gradient from hill top to bottom. Diameter at breast height (DBH) and tree height data had significant spatial auto-correlations among rows and columns. Adding a first-order separable autoregressive term more effectively modelled the spatial variation than did the incomplete block (IB) model used for the experimental design. The spatial model also accounted for effects of experimental design factors and greatly reduced residual variances. The spatial analysis relative to the IB analysis improved estimation of genetic parameters with the residual variance reduced 13 and 19% for DBH and tree height, respectively; heritability increased 35 and 51% for DBH and tree height, respectively; and genetic gain improved 3–5%. Fitting global trend and post-blocking did not improve the analyses under IB model. The use of a spatial model or combined with a design model is recommended for forest genetic trials, particularly with global trend and local spatial variation of hilly sites.

Keywords

Chinese fir / Genetic variance / Heritabilities / Progeny testing / Spatial analysis

References

[1] Bartlett MS. The approximate recovery of information from replicated field experiments with large blocks. J Agric Sci, 1938, 28: 418-427.

[2] Bian L, Shi J, Zheng R, Chen J, Wu HX. Genetic parameters and genotype–environment interactions of Chinese fir (Cunninghamia lanceolata) in Fujian province. Can J For Res, 2014, 44: 582-592.

[3] Cochran WG, Cox GM. Experimental designs, 1957 2 New York: Wiley.

[4] Costa e Silva J, Dutkowski GW, Gilmour AR. Analysis of early tree height in forest genetic trials is enhanced by including a spatially correlated residual. Can J For Res, 2001, 31: 1887-1893.

[5] Dutkowski GW, Costa e Silva J, Gilmour AR, Lopez GA. Spatial analysis methods for forest genetic trials. Can J For Res, 2002, 32: 2201-2214.

[6] Dutkowski GW, Costa e Silva J, Gilmour AR, Wellendorf H, Aguiar AM. Spatial analysis enhances modelling of a wide variety of traits in forest genetic trials. Can J For Res, 2006, 36: 1851-1870.

[7] Ericsson T. Enhanced heritabilities and best linear unbiased predictors through appropriate blocking of progeny trials. Can J For Res, 1997, 27: 2097-2101.

[8] Fang L, Shi J. Studies on comprehensive selection for superior family and clone of Chinese fir. J Nanjing For Univ, 1998, 22(1): 17-21. (Chinese with English abstract)

[9] Gilmour AR, Cullis BR. Average information reml: an efficient algorithm for variance parameter estimation in linear mixed models. Biometrics, 1995, 51: 1440-1450.

[10] Gilmour AR, Gogel BJ, Cullis BR, Thompson R. ASReml user guide release 3.0, 2009, Hemel Hempstead: VSN International Ltd.

[11] Gleeson AC, Cullis BR. Residual maximum likelihood (REML) estimation of a neighbour model for field experiments. Biometrics, 1987, 43: 277-288.

[12] Kusnandar D, Galwey N. A proposed method for estimation of genetic parameters on forest trees without raising progeny: critical evaluation and refinement. Silvae Genet, 2000, 49: 15-21.

[13] Libby WJ, Cockerham CC. Random non-contiguous plots in interlocking field layouts. Silvae Genet, 1980, 29: 183-190.

[14] Loo-Dinkins JA. Fins L, Friedman S, Brotschol J. Field test design. Handbook of quantitative forest genetics, 1992, Dordrecht: Kluwer 96 139

[15] Magnussen S. Application and comparison of spatial models in analyzing tree-genetics field trials. Can J For Res, 1990, 20: 536-546.

[16] Magnussen S. Bias in genetic variance estimates due to spatial autocorrelation. Theor Appl Genet, 1993, 86: 349-355.

[17] Martin RJ. The use of time-series models and methods in the analysis of agricultural field trials. Commun Stat Theory Methods, 1990, 19: 55-81.

[18] Papadakis JS. Méthode statistique pour des expériences sur champ. Bull Inst Amel Plantes Solonique, 1937, 23: 12-28.

[19] Patterson HD, Thompson R. Recovery of inter-block information when block sizes are unequal. Biometrika, 1971, 58: 545-554.

[20] Shi J. The Present situation of Chinese fir genetic improvement in Fujian Province and the technical countermeasures of developing it. J Fujian For Sci Technol, 1994, 21(3): 28-31. (Chinese with English abstract)

[21] Smith A, Cullis B, Thompson R. The analysis of crop cultivar breeding and evaluation trials: an overview of current mixed model approaches. J Agric Sci, 2005, 143: 449-462.

[22] White TL, Adams WT, Neale DB. Forest genetics, 2007, Wallingford: CABI Publishing

[23] Wilkinson GN, Eckert SR, Hancock TW, Mayo O. Nearest neighbor (NN) analysis of field experiments. J R Stat Soc B, 1983, 45: 151-211.

[24] Williams ER, Matheson AC, Harwood CE. Experimental design and analysis for tree improvement, 2002 2 Collingwood: CSIRO.

[25] Ye TZ, Jayawickrama KJS. Efficiency of using spatial analysis in first-generation coastal douglas-fir progeny tests in the US pacific northwest. Tree Genet Genomes, 2008, 4: 677-692.

[26] Yu R, Shi J, Liang Y, Chen X, Tan J, He Z. The analysis of open pollination progeny tests of Cunninghamia lanceolata second-generation seed orchards. J Fujian For Sci Technol, 2001, 28: 16-18. (Chinese with English abstract)