西门塔尔牛肉质性状低密度芯片的基因组选择

发布时间：2018-05-22 07:53

本文选题：全基因组关联分析 + 基因组选择　；参考：《中国农业科学院》2015年硕士论文

【摘要】：全基因组关联分析和基因组选择是近年来畜禽育种的研究热点。本研究利用BovineHD芯片,对西门塔尔牛的部分胴体性状和肉质性状进行全基因组关联分析,初步探索低密度芯片对脂肪酸含量性状的基因组预测准确性。1.使用混合压缩线性模型(CMLM)和线性模型(LM)对霖肉、后腱子和骨重三个胴体性状进行全基因组关联分析,共检验出186个显著关联的位点(P10-5),其中有55个位点在两种模型中均显著,多数标记落于6号和14号染色体的LAP3、LCORL、FAM184B、PLAG1等基因上,显著SNP重叠于相关胴体重和骨重的数量性状基因座位(QTL)。2.使用CMLM和LM对大理石花纹、脂肪颜色、总脂肪酸含量(TFA)、饱和脂肪酸含量(SFA)、单不饱和脂肪酸含量(MUFA)和多不饱和脂肪酸含量(PUFA)六个肉质性状的分析,检测出91个显著位点(P10-5),其中44个位点在两模型中均显著,与三个脂肪酸含量性状均显著相关的标记落在14号染色体的MYC基因附近,多数显著位点落在相关于大理石花纹、第十二肋的背膘厚和脂肪酸含量的QTL上。3.本研究构建不同标记数目的低密度芯片,包括均匀分布低密度芯片,基于Bayes A、Bayes B估计标记效应的绝对值及显著性的筛选标记低密度芯片,使用低密度芯片对四个脂肪酸含量性状进行基因组预测,通过五倍交叉验证衡量准确性。均匀分布低密度芯片整合BovineHD和已有低密度芯片位点,其标记数目分别为3K,7K,9K,20K和40K,准确性在9K时较高,低于筛选标记低密度芯片,与其他模拟数据中低密度芯片基因组预测的结果一致。依据标记效应及显著性的筛选标记低密度芯片,其标记数目分别为0.3 K,0.5 K,0.7 K,1 K,3 K,5 K,7 K,9 K,11 K,13 K,15 K和30K,在标记数目达到7K时准确性基本稳定,基于Bayes B估计标记效应筛选标记准确性最高,基于Bayes A估计标记效应筛选标记准确性略高于基于标记显著性筛选标记。交叉验证的对比试验中,基于同态一致性(IBS)距离矩阵分组的准确性略高于随机分组。
[Abstract]:The whole genome association analysis and genome selection are the focus of animal breeding in recent years. The whole genome association analysis of carcass traits and meat quality traits of Simmental cattle was carried out by using BovineHD chip, and the accuracy of genome prediction of fatty acid content traits by low density microarray was preliminarily explored. Using mixed compression linear model (CMLM) and linear model (LM) to analyze the whole genome association of three carcass traits, Lin-meat, posterior tendon and bone weight, a total of 186 significantly correlated loci (P10-5) were detected, 55 of which were significant in both models. Most of the markers were found on chromosome 6 and chromosome 14, such as LAP3FCL, FAM18B, PLAG1 and so on. Significant SNP overlapped with QTL1 gene locus of quantitative traits related to carcass weight and bone weight. CMLM and LM were used to analyze six fleshy characters, such as marbling, fat color, total fatty acid content, saturated fatty acid content, monounsaturated fatty acid content and polyunsaturated fatty acid content. 91 significant loci (P10-5) were detected, 44 of which were significant in both models. The markers associated with the three fatty acid content traits fell near the MYC gene on chromosome 14, and most of the significant loci were found in marbling patterns. The back fat thickness and fatty acid content of the twelfth rib were on QTL. 3. In this study, we constructed low density chips with different number of markers, including uniformly distributed low density chips, and estimated the absolute value of labeling effect and significant screening of low density chips based on Bayes Agnes Bayes B. The genome of four fatty acid content traits was predicted by low density microarray, and the accuracy was verified by five times cross validation. The uniform distribution of low density chip integrated with BovineHD and the number of low density chip sites were 3K ~ 7K ~ 9K ~ (-1) 20 K and 40 K, respectively. The accuracy of the labeled low density chip was higher at 9K than that of screening labeled low density chip. The results are consistent with those predicted by low density microarray in other simulated data. According to the labeling effect and the significance of screening low density microarray, the labeling number of the low density microarray was 0.3 KG 0.5 KX 0.7 KG 1 KN 3 KN 5 KN 7 KN 9 KN 11 KN 13 KG 15 K and 30 K, and the accuracy was stable when the labeling number reached 7 K. The accuracy of marker screening based on Bayes B estimation was the highest, and the accuracy of marker screening based on Bayes A was slightly higher than that based on marker significance. The accuracy of distance matrix grouping based on homomorphic consistency is slightly higher than that of random grouping.
【学位授予单位】：中国农业科学院
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：S823

【参考文献】