20317

Orman ağacı ıslah çalışmalarında doğrusal karma model seçimi

Orman ağacı ıslah çalışmalarında, uzun süre gözlemlenen genetik testler ile ıslah programları için genetik parametreler tahmin edilmektedir. Söz konusu parametrelerin tahminleri ıslah programını etkileyeceğinden tahmin için kullanılacak doğrusal karma modelin seçimi büyük önem taşımaktadır. Kullanılan doğrusal karma modellerde tahmin, genellikle artık (residual) veya kısıtlı maksimum olabilirlik (REML) yöntemi kullanılarak elde edilir. Farklı sabit etkileri olan modellerin olabilirliğe (likelihood) dayalı bilgi kriterleri ile kıyaslanabilmesi için, modellerin maksimum olabilirlik (maximum likelihood) kullanılarak tahmin edilmesi önerilmektedir. Orman ağaçları ıslahı çalışmalarında doğrusal karma model seçiminde farklı modeller denenerek model uyumunu arttıran en kullanışlı model seçilmelidir. Orman ağaçları ıslah çalışmalarında modellerin uyumunu kıyaslamak için ise genellikle Akaike (AIC) bilgi kriterinin kullanılması önerilmektedir. Bu çalışmada, doğrusal karma model seçiminin gerekliliğini ve önemini ortaya koymak amaçlanmıştır. Bu amaç için, Muğla-Marmaris’te açık tozlaşma ürünü 168 aile (üvey kardeş) ile tesis edilmiş olan Kızılçam döl deneme sahasındaki ağaçların on ikinci yaş göğüs yüksekliği çap verileri kullanılarak modeller kıyaslanmıştır. Verilerin analizinde geleneksel (basit), mekânsal bileşen içeren, artığın bağımsız veya birinci dereceden iki boyutlu ayrılabilir otoregresif korelasyon hata yapısı olduğunu varsayan toplamda otuziki farklı model denenmiştir. Geleneksel modelin AIC değeri (Model-1=5594.1), mekânsal bileşen ve artığın otoregresif korelasyon yapısı içeren modellere kıyasla (Model-20=5447) daha yüksek bulunmuştur

Anahtar Kelimeler:

Model Seçimi, Doğrusal Karma Modeller, Mekansal Analiz, Akaike Bilgi Kriteri, Döl Denemeleri, Kızılçam

Linear mixed model selection in forest tree breeding studies

In forest tree breeding studies, genetic parameters are estimated for breeding programs with long-term genetic tests. Since the estimations of these parameters will affect the breeding program, the selection of the linear mixed model to be used for the estimation is of great importance. In linear mixed models used for the estimation of these parameters, the estimation is usually obtained by using the residual or restricted maximum likelihood (REML) methods. In order to compare the models with different fixed effects with the information criteria based on likelihood, it is suggested that the models should be estimated using maximum likelihood. In forest tree breeding studies, different models should be tried in linear mixed model selection and the most useful model that increases model fit should be selected. It is generally recommended to use the Akaike (AIC) information criterion to compare the fit of models in forest tree breeding studies. In this study, it is aimed to reveal the necessity and importance of linear mixed model selection. For this purpose, the models were compared by using the twelfth year diameter at breast height data of the trees in the red pine progeny trial field established with 168 open pollinated families (half-sib) in Muğla-Marmaris. In the analysis of this data set, a total of 32 different models were tested, which include traditional (simple), spatial component, and assume that the residual is an independent or 1st order 2-dimensional separable autoregressive correlation error structure. The AIC value of the traditional model (Model-1=5594.1) was found to be higher than the models with the spatial component and the residual autoregressive correlation structure (Model-20=5447).

Keywords:

Model selection, Linear mixed models, Spatial Analysis, Akaike ınformation criteria, Progeny trials, Turkish Red Pine,

PDF

___

Alan M, Öztürk H, Şıklar S, Ezen T, Korkmaz B, Doğan B, Keskin S, Tulukçu M, Derilgen SI, Çalışkan B (2005) Ege bölgesi alt yükselti kuşağı ıslah zonunda (0-400 m) Kızılçam (Pinus brutia ten.) döl denemeleri 4. yaş sonuçları. (Teknik Bülten:13), Orman Ağaçları ve Tohumları Islah Araştırma Enstitüsü Müdürlüğü, Ankara.
Aparicio J, Ariza-Suarez D, Raatz B (2019) Web application for spatial modelling of field trials. Paper presented at the XXIX Simposio Internacional de Estadística, Barranquilla
Borges A, González-Reymundez A, Ernst O, Cadenazzi M, Terra J, Gutiérrez L (2019) Can spatial modeling substitute for experimental design in agricultural experiments? Crop Science, 59(1):44-53
Bozdogan H (1987) Model selection and Akaike's information criterion (AIC): The general theory and its analytical extensions. Psychometrika, 52(3):345-370
Brown H, Prescott R (2015) Applied mixed models in medicine. John Wiley & Sons, New York
Burgueño J, Cadena A, Crossa J, Banziger M, Gilmour A, Cullis B (2000) User's guide for spatial analysis of field variety trials using ASREML. International Maize and Wheat Improvement Center (CIMMYT), Mexico
Caliński T, Czajka S, Kaczmarek Z, Krajewski P, Pilarczyk W (2005) Analyzing multi‐environment variety trials using randomization‐derived mixed models. Biometrics, 61(2):448-455
Casler MD (2015) Fundamentals of experimental design: Guidelines for designing successful experiments. Agronomy Journal, 107(2):692-705
Chen Z, Karlsson B, Wu H (2017) Patterns of additive genotype-by environment interaction in tree height of Norway spruce in southern and central Sweden. Tree Genetics & Genomes, 13(1):1-14
Cullis BR, Smith AB, Coombes NE (2006) On the design of early generation variety trials with correlated data. Journal of Agricultural, Biological, Environmental Statistics, 11(4):381-393
Dutkowski GW, Costa e Silva J, Gilmour AR, Wellendorf H, Aguiar A (2006) Spatial analysis enhances modelling of a wide variety of traits in forest genetic trials. Canadian Journal of Forest Research, 36(7):1851-1870
Dutkowski GW, Silva JCe, Gilmour AR, Lopez GA (2002) Spatial analysis methods for forest genetic trials. Canadian Journal of Forest Research, 32(12):2201-2214
Fisher RA (1936) Design of experiments. British Medical Journal 1(3923):554
Galwey NW (2014) Introduction to mixed modelling: beyond regression and analysis of variance. John Wiley & Sons, New York
Gezan SA, White TL, Huber DA (2010) Accounting for spatial variability in breeding trials: a simulation study. Agronomy Journal, 102(6):1562-1571
Gilmour A, Gogel B, Cullis B, Thompson R (2009) ASReml user guide release 3.0 VSN International Ltd, Hemel Hempstead
Gilmour A, Gogel B, Cullis B, Welham S, Thompson R (2015) ASReml user guide release 4.1 functional specification. VSN International Ltd, Hemel Hempstead
Grondona M, Crossa J, Fox P, Pfeiffer W (1996) Analysis of variety yield trials using two-dimensional separable ARIMA processes. Biometrics, 763-770
Guerin L, Stroup WW (2000) A simulation study to evaluate PROC MIXED analysis of repeated measures data Isik F, Holland J, Maltecca C (2017) Genetic data analysis for plant and animal breeding. Springer International Publishing, Switzerland
Kehel Z, Habash D, Gezan S, Welham S, Nachit MJAJ (2010) Estimation of spatial trend and automatic model selection in augmented designs. Agronomy Journal, 102(6):1542-1552
Legendre P (1993) Spatial autocorrelation: trouble or new paradigm? Ecology, 74(6):1659-1673
Little R, Milliken GA, Stroup WW, Wolfinger RD, Schabenberger O (2006) SAS for mixed models. SAS Institute Inc, North Carolina
Matheson A, Gapare W, Ilic J, Wu H (2008) Inheritance and genetic gain in wood stiffness in radiata pine assessed acoustically in young standing trees. Silvae Genetica, 57(2):56-64
Mathew B, Holand AM, Koistinen P, Léon J, Sillanpää MJ (2016) Reparametrization-based estimation of genetic parameters in multi-trait animal model using Integrated Nested Laplace Approximation. Theoretical Applied Genetics, 129(2):215-225
Mramba LK (2016) Optimal experimental designs for spatially and genetically correlated data using linear mixed models. University of Florida
Piepho HP, Möhring J, Melchinger AE, Büchse A (2008) BLUP for phenotypic selection in plant breeding and variety testing. Euphytica, 161(1-2):209-228
Piepho HP, Williams ER, Michel V (2015) Beyond latin squares: a brief tour of row‐column designs. Agronomy Journal, 107(6):2263-2270 Schutz W, Cockerham CC (1966) The effect of field blocking on gain from selection. Biometrics, 22(4):843-863 Shalizi MN, Gezan SA, McKeand SE, Sherrill JR, Cumbie WP, Whetten RW, Isik F (2020) Correspondence between breeding values of the same Pinus taeda L. genotypes from clonal trials and half-sib seedling progeny trials. Forest Science, 66(5):600-611
Smith A, Cullis B, Thompson R (2001) Analyzing variety by environment data using multiplicative mixed models and adjustments for spatial field trend. Biometrics, 57(4):1138-1147
Spilke J, Richter C, Piepho HJPb (2010) Model selection and its consequences for different split‐plot designs with spatial covariance and trend. Plant Breeding, 129(6):590-598
Stroup WW (2012) Generalized linear mixed models: modern concepts, methods and applications. CRC Press, Boca Raton.
Urhan OS, Kolpak SE, Jayawickrama KJS, Howe GT (2014) Early genetic selection for wood stiffness in juvenile Douglas-fir and western hemlock. Forest Ecology and Management, 320:104-117
Verbyla AP (2019) A note on model selection using information criteria for general linear models estimated using REML. Australian & New Zealand Journal of Statistics, 61(1):39-50
VSNI T (2021) A brief look at spatial modelling. In. /, (01.10.2022).
Welham SJ, Gezan SA, Clark SJ, Mead A (2014) Statistical methods in biology: design and analysis of experiments and regression. CRC Press, London.
Welham SJ, Gogel BJ, Smith AB, Thompson R, Cullis BR (2010) A comparison of analysis methods for late‐stage variety evaluation trials. Australian and New Zealand Journal of Statistics, 52(2):125-149
White TL, Adams WT, Neale DB (2007) Forest genetics. Cabi, United Kingdom.
Yates F (1939) The recovery of inter‐block information in variety trials arranged in three‐dimensional lattices. Annals of Eugenics, 9(2):136-156
Ye TZ, Jayawickrama KJ (2008) Efficiency of using spatial analysis in first-generation coastal Douglas-fir progeny tests in the US Pacific Northwest. Tree Genetics & Genomes, 4(4):677-692