進階回交族群之數量性狀基因座定位門檻值研究

A study of assessing genome-wise statistical significance for QTL mapping in the advanced backcross populations

郭又慈(You-Tsz Guo)

指導教授 ： 高振宏

共同指導教授 ： 廖振鐸

國立臺灣大學/生物資源暨農學院/農藝學研究所/碩士(2011年)

https://doi.org/10.6342/NTU.2011.03012

摘要

數量性狀基因座（quantitative trait loci, QTL）區間定位法的統計模式一般是混合常態模型。在偵測QTL時，我們使用概度比檢定統計量來檢定整個基因組的每一個位置上是否有QTL存在，其虛無假設為QTL不存在，而具有顯著最大概度比檢定統計量之位置即為QTL的估計位置。在這樣的框架之下，決定偵測QTL時所宣稱之顯著性的門檻值，在QTL定位上是一個非常重要且具有挑戰性的議題。目前為止，有關決定門檻值的研究大多在回交族群及F2世代中進行。在實際的植物和動物育種研究裡，回交及F2世代的進階族群也很常被使用，而這些族群有很不同的基因組結構。在這項研究中，我們使用score檢定統計量和高斯隨機過程來取得進階回交族群之門檻值，並研究他們在進階回交族群之QTL定位時的行為。使用此方法我們需要考慮這些進階族群的特定基因組結構，因此我們推導出三個基因及四個基因的傳遞方程式（transition equations）來計算得到基因型分佈，並將這些基因型頻率帶入score檢定統計量和高斯隨機過程的公式，來計算取得不同族群大致的QTL定位門檻值。進行模擬來印證我們的方法。

關鍵字

數量性狀基因座 ； 區間定位法 ； 進階回交族群 ； score檢定 ； 高斯隨機過程

並列摘要

The statistical model of interval mapping for QTL (quantitative trait loci) detection is generally a normal mixture model. In detecting QTL, typically the presence of a QTL, i.e. the null hypothesis of no QTL, is tested over the all possible positions in the whole genome by using likelihood ratio test (LRT) statistics and the position with the maximum significant LRT statistic is regarded as the estimated QTL position. Under such a framework, the determination of the threshold values for declaring the significance of QTL detection has been recognized as a very important and challenging issue in QTL mapping. So far, most of the studies related to determining the threshold values are performed for the backcross and F$_2$ populations. In practical plant and animal breeding studies, advanced populations from backcross or F2 populations, which can have very different genome structures, are also very popular. In this study, we use score test statistics and Gaussian stochastic process to obtain the threshold values and investigate their behaviors for QTL mapping in the advanced backcross populations. To consider the specific genome structures of the advanced populations in the approach,
we derive the sets of transition equations to obtain the genotypic distributions of three and four genes and devise these genotypic frequencies into the formulations of the score test statistics and Gaussian processes to compute the approximate threshold values for different populations.
Simulation studies are performed to verify our approach.

並列關鍵字

QTL ； interval mapping ； advanced backcross popultions ； score test ； Gaussian stochastic process

參考文獻

Kao, Chen-Hung (2004), “Multiple-interval mapping for quantitative trait loci controlling endosperm traits”, Genetics, 167(4), 1987–2002.

Bernacchi, D., Beck-Bunn, T., Eshed, Y., Lopez, J., Petiard, V., Uhlig, J., Zamir, D., and Tanksley, S. (1998), “Advanced backcross qtl analysis in tomato. i. identification of qtls for traits of agronomic importance from lycopersicon hirsutum”, TAG Theoretical and Applied Genetics, 97, 381–397, 10.1007/s001220050908.

Chang, Myron N., Wu, Rongling, Wu, Samuel S., and Casella, George (2009), “Score statistics for mapping quantitative trait loci”, Statistical Applications in Genetics and Molecular Biology, 8(1).

Churchill, G. A. and Doerge, R. W. (1994), “Empirical threshold values for quantitative trait mapping”, Genetics, 138(3), 963–971.

Davies, ROBERT B. (1977), “Hypothesis testing when a nuisance parameter is present only under the alternative”, Biometrika, 64(2), 247–254.