首页 > 其他分享> > 使用TASSEL学习GWAS笔记（6/6）：TASSEL结果可视化：QQ plot，曼哈顿图

使用TASSEL学习GWAS笔记（6/6）：TASSEL结果可视化：QQ plot，曼哈顿图

2021-10-07 20:29:57 作者：互联网

笔记计划分为六篇：

第一篇：读取plink基因型数据和表型数据
 第二篇：对基因型数据质控：缺失质控，maf质控，hwe质控，样本质控
 第三篇：基因型数据可视化：kingship，LD，MDS，PCA
第四篇：一般线性模型进行GWAS分析（GLM模型）
第五篇：混合线性模型进行GWAS分析（MLM模型）
第六篇：TASSEL结果可视化：QQ plot，曼哈顿图

已完成前五篇，本篇是第六篇。

1. TASSEL的GLM和MLM分析结果

质控后的plink数据和表型数据：

GLM的GWAS分析结果：

MLM的GWAS分析结果：

2. TASSEL中的可视化

TASSEL有对结果进行可视化的模块，包括qq图和曼哈顿图，但是图不方便调整。这里用TASSEL的分析结果，使用R语言进行绘制qq图和曼哈顿图。

3. R语言包安装及载入

需要用到：

qqman
tidyverse
data.table

下面代码，会判断是否有这三个包，如果没有，就自动安装。然后载入软件包。

if(!require(data.table)) install.packages("data.table")
if(!require(qqman)) install.packages("qqman")
if(!require(tidyverse)) install.packages("tidyverse")

library(qqman)
library(tidyverse)
library(data.table)

4. GLM模型GWAS结果可视化

results_log = fread("glm-result.txt")
dim(results_log)
head(results_log)

select = dplyr::select
table(results_log$Trait)

结果：

> table(results_log$Trait)

 dpoll EarDia  EarHT 
  2460   2460   2460

数据中共有三个性状，可以选择一个性状，进行可视化。

d1 = results_log %>% filter(Trait == "dpoll") %>% select(Chr,Marker,Pos,p)
head(d1)
summary(d1)
d1 = d1 %>% drop_na(p)
summary(d1)

注意，有些P值是NA，在作图时会报错，这里将其移除。

整理后的结果：

> summary(d1)
      Chr          Marker               Pos                  p         
 Min.   : 1.0   Length:2460        Min.   :   139753   Min.   :0.0000  
 1st Qu.: 2.0   Class :character   1st Qu.: 43868061   1st Qu.:0.1236  
 Median : 4.0   Mode  :character   Median :128423374   Median :0.3911  
 Mean   : 4.7                      Mean   :120382976   Mean   :0.4165  
 3rd Qu.: 7.0                      3rd Qu.:175628840   3rd Qu.:0.6743  
 Max.   :10.0                      Max.   :298413352   Max.   :0.9996

作图代码：

manhattan(d1,chr="Chr",bp="Pos",p="p",snp="Marker", main = "Manhattan plot: logistic")
tiff("y1-曼哈顿图.tiff")
manhattan(d1,chr="Chr",bp="Pos",p="p",snp="Marker", main = "Manhattan plot: logistic")
dev.off()

qq(d1$p, main = "Q-Q plot of GWAS p-values : log")
tiff("y1-QQ图.tiff")
qq(d1$p, main = "Q-Q plot of GWAS p-values : log")
dev.off()

曼哈顿图：

QQ图：

其它两个性状的作图代码：


d2 = results_log %>% filter(Trait == "EarDia") %>% select(Chr,Marker,Pos,p)
head(d2)
summary(d2)
d2 = d2 %>% drop_na(p)
summary(d2)

tiff("y2-曼哈顿图.tiff")
manhattan(d2,chr="Chr",bp="Pos",p="p",snp="Marker", main = "Manhattan plot: logistic")
dev.off()

tiff("y2-QQ图.tiff")
qq(d2$p, main = "Q-Q plot of GWAS p-values : log")
dev.off()


d3 = results_log %>% filter(Trait == "EarHT") %>% select(Chr,Marker,Pos,p)
head(d3)
summary(d3)
d3 = d3 %>% drop_na(p)
summary(d3)

tiff("y3-曼哈顿图.tiff")
manhattan(d3,chr="Chr",bp="Pos",p="p",snp="Marker", main = "Manhattan plot: logistic")
dev.off()

tiff("y3-QQ图.tiff")
qq(d3$p, main = "Q-Q plot of GWAS p-values : log")
dev.off()

将整理后的不同性状的结果保存到本地：

fwrite(d1,"y1_result.csv")
fwrite(d2,"y2_result.csv")
fwrite(d3,"y3_result.csv")

5. MLM模型GWAS结果可视化

读取数据，提取性状，去掉P值为缺失的行：

library(qqman)
library(data.table)
results_log = fread("mlm-result.txt", head=TRUE)
dim(results_log)
head(results_log)

library(tidyverse)
select = dplyr::select
table(results_log$Trait)
d1 = results_log %>% filter(Trait == "dpoll") %>% select(Chr,Marker,Pos,p)
head(d1)
summary(d1)
d1 = d1 %>% drop_na(p)
summary(d1)

曼哈顿图：

QQ图：

其它两个作图代码：


d2 = results_log %>% filter(Trait == "EarDia") %>% select(Chr,Marker,Pos,p)
head(d2)
summary(d2)
d2 = d2 %>% drop_na(p)
summary(d2)

tiff("y2-曼哈顿图.tiff")
manhattan(d2,chr="Chr",bp="Pos",p="p",snp="Marker", main = "Manhattan plot: logistic")
dev.off()

tiff("y2-QQ图.tiff")
qq(d2$p, main = "Q-Q plot of GWAS p-values : log")
dev.off()


d3 = results_log %>% filter(Trait == "EarHT") %>% select(Chr,Marker,Pos,p)
head(d3)
summary(d3)
d3 = d3 %>% drop_na(p)
summary(d3)

tiff("y3-曼哈顿图.tiff")
manhattan(d3,chr="Chr",bp="Pos",p="p",snp="Marker", main = "Manhattan plot: logistic")
dev.off()

tiff("y3-QQ图.tiff")
qq(d3$p, main = "Q-Q plot of GWAS p-values : log")
dev.off()

6. 完整代码汇总

GLM的可视化代码：

## 对TASSEL GLM 模型可视化

if(!require(data.table)) install.packages("data.table")
if(!require(qqman)) install.packages("qqman")
if(!require(tidyverse)) install.packages("tidyverse")

library(qqman)
library(tidyverse)
library(data.table)

results_log = fread("glm-result.txt")
dim(results_log)
head(results_log)

select = dplyr::select
table(results_log$Trait)
d1 = results_log %>% filter(Trait == "dpoll") %>% select(Chr,Marker,Pos,p)
head(d1)
summary(d1)
d1 = d1 %>% drop_na(p)
summary(d1)

manhattan(d1,chr="Chr",bp="Pos",p="p",snp="Marker", main = "Manhattan plot: logistic")
tiff("y1-曼哈顿图.tiff")
manhattan(d1,chr="Chr",bp="Pos",p="p",snp="Marker", main = "Manhattan plot: logistic")
dev.off()

qq(d1$p, main = "Q-Q plot of GWAS p-values : log")
tiff("y1-QQ图.tiff")
qq(d1$p, main = "Q-Q plot of GWAS p-values : log")
dev.off()

d2 = results_log %>% filter(Trait == "EarDia") %>% select(Chr,Marker,Pos,p)
head(d2)
summary(d2)
d2 = d2 %>% drop_na(p)
summary(d2)

tiff("y2-曼哈顿图.tiff")
manhattan(d2,chr="Chr",bp="Pos",p="p",snp="Marker", main = "Manhattan plot: logistic")
dev.off()

tiff("y2-QQ图.tiff")
qq(d2$p, main = "Q-Q plot of GWAS p-values : log")
dev.off()


d3 = results_log %>% filter(Trait == "EarHT") %>% select(Chr,Marker,Pos,p)
head(d3)
summary(d3)
d3 = d3 %>% drop_na(p)
summary(d3)

tiff("y3-曼哈顿图.tiff")
manhattan(d3,chr="Chr",bp="Pos",p="p",snp="Marker", main = "Manhattan plot: logistic")
dev.off()

tiff("y3-QQ图.tiff")
qq(d3$p, main = "Q-Q plot of GWAS p-values : log")
dev.off()

fwrite(d1,"y1_result.csv")
fwrite(d2,"y2_result.csv")
fwrite(d3,"y3_result.csv")

MLM的可视化代码：

## 对TASSEL GLM 模型可视化

library(qqman)
library(data.table)
results_log = fread("mlm-result.txt", head=TRUE)
dim(results_log)
head(results_log)

library(tidyverse)
select = dplyr::select
table(results_log$Trait)
d1 = results_log %>% filter(Trait == "dpoll") %>% select(Chr,Marker,Pos,p)
head(d1)
summary(d1)
d1 = d1 %>% drop_na(p)
summary(d1)

manhattan(d1,chr="Chr",bp="Pos",p="p",snp="Marker", main = "Manhattan plot: logistic")
tiff("y1-曼哈顿图.tiff")
manhattan(d1,chr="Chr",bp="Pos",p="p",snp="Marker", main = "Manhattan plot: logistic")
dev.off()

qq(d1$p, main = "Q-Q plot of GWAS p-values : log")
tiff("y1-QQ图.tiff")
qq(d1$p, main = "Q-Q plot of GWAS p-values : log")
dev.off()

d2 = results_log %>% filter(Trait == "EarDia") %>% select(Chr,Marker,Pos,p)
head(d2)
summary(d2)
d2 = d2 %>% drop_na(p)
summary(d2)

tiff("y2-曼哈顿图.tiff")
manhattan(d2,chr="Chr",bp="Pos",p="p",snp="Marker", main = "Manhattan plot: logistic")
dev.off()

tiff("y2-QQ图.tiff")
qq(d2$p, main = "Q-Q plot of GWAS p-values : log")
dev.off()


d3 = results_log %>% filter(Trait == "EarHT") %>% select(Chr,Marker,Pos,p)
head(d3)
summary(d3)
d3 = d3 %>% drop_na(p)
summary(d3)

tiff("y3-曼哈顿图.tiff")
manhattan(d3,chr="Chr",bp="Pos",p="p",snp="Marker", main = "Manhattan plot: logistic")
dev.off()

tiff("y3-QQ图.tiff")
qq(d3$p, main = "Q-Q plot of GWAS p-values : log")
dev.off()

fwrite(d1,"y1_result.csv")
fwrite(d2,"y2_result.csv")
fwrite(d3,"y3_result.csv")

标签：QQ,plot,TASSEL,log,tiff,d2,d3,d1
来源： https://blog.csdn.net/yijiaobani/article/details/119885686