Comparison of statistical models for nested association mapping in rapeseed (Brassica napus L.) through computer simulations

BMC Plant Biology

Research Article

Benjamin Stich¹  Anja Bus¹  Jinquan Li¹  Viola Spamer² 

[1] Max Planck Institute for Plant Breeding Research, Carl-von-Linné-Weg 10, 50829, Köln, Germany;Syngenta Seeds GmbH, Zum Knipkenbach 20, 32107, Bad Salzuflen, Germany;

关键词: Statistical models;    Nested association mapping (NAM);    Brassica napus;    Double haploid NAM;    Backcross NAM;    Computer simulations;   

DOI  :  10.1186/s12870-016-0707-6

 received in 2015-04-21, accepted in 2016-01-07,  发布年份 2016

来源: Springer

【 摘 要 】

BackgroundRapeseed (Brassica napus L.) is an important oilseed crop throughout the world, serving as source for edible oil and renewable energy. Development of nested association mapping (NAM) population and methods is of importance for quantitative trait locus (QTL) mapping in rapeseed. The objectives of the research were to compare the power of QTL detection 1- β∗ (β∗ is the empirical type II error rate) (i) of two mating designs, double haploid (DH-NAM) and backcross (BC-NAM), (ii) of different statistical models, and (iii) for different genetic situations.ResultsThe computer simulations were based on the empirical data of a single nucleotide polymorphism (SNP) set of 790 SNPs from 30 sequenced conserved genes of 51 accessions of world-wide diverse B. napus germplasm. The results showed that a joint composite interval mapping (JCIM) model had significantly higher power of QTL detection than a single marker model. The DH-NAM mating design showed a slightly higher power of QTL detection than the BC-NAM mating design. The JCIM model considering QTL effects nested within subpopulations showed higher power of QTL detection than the JCIM model considering QTL effects across subpopulations, when examing a scenario in which there were interaction effects by a few QTLs interacting with a few background markers as well as a scenario in which there were interaction effects by many QTLs (≥25\documentclass[12pt]{minimal}\usepackage{amsmath}\usepackage{wasysym}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{amsbsy}\usepackage{mathrsfs}\usepackage{upgreek}\setlength{\oddsidemargin}{-69pt}\begin{document}$\geqslant 25$\end{document}) each with more than 10 background markers and the proportion of total variance explained by the interactions was higher than 75 %.ConclusionsThe results of our study support the optimal design as well as analysis of NAM populations, especially in rapeseed.

【 授权许可】

CC BY   
© Li et al. 2016

【 图 表 】

Fig. 5

Fig. 4

12936_2016_1411_Article_IEq110.gif

Fig. 2

12888_2023_5292_Article_IEq1.gif

Fig. 1

Fig. 2