博文
阈值性状与连续性状&阈值性状与阈值性状的遗传相关分析-asrmel4r
||
阈值性状与连续性状&阈值性状与阈值性状的遗传相关分析-asrmel4r
asreml3中，还不支持阈值性状与连续性状的相关分析，现在asreml4r中支持了，下面用实际的案例介绍一下实现方法
与两性状分析的模型基本一致，关键是指定两性状的分布：family=list(asr_binomial(),asr_gaussian())
library(asreml)
Loading required package: Matrix
查看数据
data("binnor")
head(binnor)| record | year | Grp | Sex | Sire | score5 | score4 | scald | rot | norm |
|---|---|---|---|---|---|---|---|---|---|
| 1 | 1 | 1 | 1 | 1 | 0 | 1 | 1 | 1 | 11.40431 |
| 2 | 1 | 1 | 1 | 1 | 0 | 1 | 1 | 0 | 11.38088 |
| 3 | 1 | 1 | 1 | 1 | 0 | 1 | 1 | 0 | 10.89160 |
| 4 | 1 | 1 | 1 | 1 | 0 | 1 | 1 | 0 | 11.33378 |
| 5 | 1 | 1 | 1 | 1 | 0 | 1 | 1 | 0 | 11.21983 |
| 6 | 1 | 1 | 1 | 1 | 0 | 1 | 1 | 0 | 11.38313 |
查看阈值数据：0,1
hist(binnor$score4);hist(binnor$rot)
查看连续性状
hist(binnor$norm)
阈值性状与连续性状相关分析
binnor.as <- asreml(cbind(score4,rot) ~ trait:Sex + trait:Grp, random = ~ corgh(trait):Sire, family = list(asr_binomial(),asr_binomial()), maxit=30, data=binnor)
Model fitted using the sigma parameterization. ASReml 4.1.0 Tue Mar 20 10:41:12 2018 Binomial; Logit Mu=P=1/(1+exp(-XB)); V=Mu(1-Mu)/N Note: The LogLik value is unsuitable for comparing GLM models LogLik Sigma2 DF wall cpu 1 -8608.147 1.0 5014 10:41:13 0.0 (2 restrained) 2 -6006.509 1.0 5014 10:41:13 0.0 (2 restrained) 3 -3180.068 1.0 5014 10:41:13 0.0 4 -1256.398 1.0 5014 10:41:13 0.0 5 -239.576 1.0 5014 10:41:13 0.0 6 55.568 1.0 5014 10:41:13 0.0 7 107.915 1.0 5014 10:41:13 0.0 8 112.680 1.0 5014 10:41:13 0.0 9 113.746 1.0 5014 10:41:13 0.0 10 114.007 1.0 5014 10:41:13 0.0 11 114.066 1.0 5014 10:41:13 0.0 12 114.077 1.0 5014 10:41:13 0.0 13 114.076 1.0 5014 10:41:13 0.0 14 114.074 1.0 5014 10:41:13 0.0 Deviance from GLM fit: 2971.07 Variance heterogenity factor (Deviance/df): 0.59 (assuming 5014 degrees of freedom)
summary(binnor.as)$varcomp
| component | std.error | z.ratio | bound | %ch | |
|---|---|---|---|---|---|
| trait:Sire!trait!rot:!trait!score4.cor | -0.7246951818 | 0.4167573139 | -1.738890 | U | 0.1 |
| trait:Sire!trait_score4 | 0.0340338029 | 0.0283167552 | 1.201896 | P | 0.1 |
| trait:Sire!trait_rot | 0.0003618679 | 0.0002425170 | 1.492134 | P | 0.0 |
| units:trait!R | 1.0000000000 | NA | NA | F | 0.0 |
| units:trait!trait_score4:score4 | 0.9993337451 | 0.0277407443 | 36.024042 | P | 0.0 |
| units:trait!trait_rot:score4 | 0.0412169466 | 0.0021105084 | 19.529392 | P | 0.1 |
| units:trait!trait_rot:rot | 0.0255003677 | 0.0007201911 | 35.407782 | P | 0.0 |
可以看出，score4和norm的遗传相关为-0.72，标准误为0.41
阈值性状与阈值性状相关分析
binnor.as <- asreml(cbind(score4,rot) ~ trait:Sex + trait:Grp, random = ~ corgh(trait):Sire, family = list(asr_binomial(),asr_binomial()), maxit=30, data=binnor)
Model fitted using the sigma parameterization. ASReml 4.1.0 Tue Mar 20 10:43:53 2018 Binomial; Logit Mu=P=1/(1+exp(-XB)); V=Mu(1-Mu)/N Note: The LogLik value is unsuitable for comparing GLM models LogLik Sigma2 DF wall cpu 1 -8608.147 1.0 5014 10:43:53 0.0 (2 restrained) 2 -6006.509 1.0 5014 10:43:53 0.0 (2 restrained) 3 -3180.068 1.0 5014 10:43:53 0.0 4 -1256.398 1.0 5014 10:43:53 0.0 5 -239.576 1.0 5014 10:43:53 0.0 6 55.568 1.0 5014 10:43:53 0.0 7 107.915 1.0 5014 10:43:53 0.0 8 112.680 1.0 5014 10:43:53 0.0 9 113.746 1.0 5014 10:43:53 0.0 10 114.007 1.0 5014 10:43:53 0.0 11 114.066 1.0 5014 10:43:53 0.0 12 114.077 1.0 5014 10:43:53 0.0 13 114.076 1.0 5014 10:43:53 0.0 14 114.074 1.0 5014 10:43:53 0.0 Deviance from GLM fit: 2971.07 Variance heterogenity factor (Deviance/df): 0.59 (assuming 5014 degrees of freedom)
summary(binnor.as)$varcomp
| component | std.error | z.ratio | bound | %ch | |
|---|---|---|---|---|---|
| trait:Sire!trait!rot:!trait!score4.cor | -0.7246951818 | 0.4167573139 | -1.738890 | U | 0.1 |
| trait:Sire!trait_score4 | 0.0340338029 | 0.0283167552 | 1.201896 | P | 0.1 |
| trait:Sire!trait_rot | 0.0003618679 | 0.0002425170 | 1.492134 | P | 0.0 |
| units:trait!R | 1.0000000000 | NA | NA | F | 0.0 |
| units:trait!trait_score4:score4 | 0.9993337451 | 0.0277407443 | 36.024042 | P | 0.0 |
| units:trait!trait_rot:score4 | 0.0412169466 | 0.0021105084 | 19.529392 | P | 0.1 |
| units:trait!trait_rot:rot | 0.0255003677 | 0.0007201911 | 35.407782 | P | 0.0 |
可以看出，rot和score4的遗传相关为-0.72，标准误为0.41
https://blog.sciencenet.cn/blog-2577109-1104818.html
上一篇：如何使用GGE模型分析品种稳定性
下一篇：数量遗传学理论框架下的进化生态学研究： 以动物模型的运用为例
