| 摘要: |
| 【目的】基于协方差估计的多因变量回归（multivariate regression with covariance estimation,MRCE）模型进行多性状QTL定位分析，为动植物数量性状基因定位提供理论参考。【方法】构建适用QTL定位的MRCE模型，设计3个模拟试验对模型进行检验，通过计算机生成基因型和2个相关性状的表型值，并用2组数据对模型进行实际应用，其中一组为水稻DH群体数据，选自qtlnetwork软件；另一组为水稻永久F2群体数据，由珍汕97×明恢63，含有210个株系的重组自交系（RIL）群体随机交配生成，分析MRCE模型在以上2组数据多性状QTL定位中的应用效果。【结果】用MRCE模型进行QTL定位的模拟试验结果表明，遗传变异所占方差比越大，相关系数绝对值越大，遗传率越大，则功效越好，估计值越接近效应值。MRCE的QTL定位应用结果显示，从水稻DH群体中识别出8个QTL与ph6性状有关，6个QTL与ph8性状有关；从1998年水稻永久F2群体数据中识别出3个QTL与穗粒数相关，10个QTL与粒质量相关；从1999年数据识别出3个QTL与穗粒数相关，6个QTL与粒质量相关。【结论】利用MRCE模型进行多性状QTL定位是可行的。 |
| 关键词: 数量性状 QTL定位 多因变量 协方差估计 回归模型 |
| DOI： |
| 分类号: |
| 基金项目:国家自然科学基金项目（32071892）；福建省自然科学基金项目（2021J01126）；福建农林大学科技创新专项基金项目（CXZX2019127G） |
|
| QTL mapping analysis of multiple quantitative traits based on multivariate regression with covariance estimation |
|
ZHANG Hui,YE Jingshan,SHEN Jiayu,et al
|
| Abstract: |
| 【Objective】A QTL mapping method of multiple traits was proposed based on multivariate regression with covariance estimation (MRCE) to provide reference for mapping quantitative traits of animals and plants in practice.【Method】The genetic model by MRCE was constructed for QTL mapping and validated by three simulations.Genotype and 2 related phenotype values were generated by computer simulation.The data of DH population in rice were selected from qtlnetwork software for first application example.The immortalized F2 population of rice was generated by random hybridization of a recombinant inbred line population (210 lines) from Zhenshan 97×Minghui 63 for second application example.Then,the two examples were used to analyze the application of MRCE in QTL mapping analysis on multiple quantitative traits.【Result】The QTL mapping by MRCE showed that the power of QTL detection increased and the estimation accuracies increased as the increase of genetic variant effect of variance,absolute value of correlation coefficient of phenotype and QTL heritability.For first application example,8 QTLs were identified for ph6 traits and 6 QTLs were related to ph8 traits for DH population of rice.For second application example,3 QTLs were identified for grains per panicle and 10 QTLs were identified for grain weight by the data of immortalized F2 population of rice in 1998.3 QTLs were identified for grains per panicle and 6 QTLs were related to grain weight by the data of immortalized F2 population of rice in 1999.【Conclusion】It was feasible to use MRCE for QTL mapping of multiple quantitative traits. |
| Key words: quantitative trait QTL mapping multiple-dependent variables covariance estimation regression model |
