vcf数据是保存变异信息的主要数据格式，plink是进行全基因组关联分析（GWAs）分析的常用工具包，同时提供一系列数据转换、裁剪和遗传统计量计算工具。本文以实际数据提供基因组关联分析方法。

1 数据准备

首先，使用plink将原始的SNP数据（snps_fil.recode.vcf ）转换为二进制数据（bim、bed、fam）：
格式转换：

mkdir GWAs_raw
nohup plink --vcf snps_fil.recode.vcf --threads 20 --const-fid --allow-extra-chr --make-bed -out GWAs_raw/snps_fil >> 20231111_vcf_to_bimbedfam.log&

将包含有亲本信息，详细介绍：
更改FID：

plink --bfile GWAs_raw/snps_fil --update-ids POP.FID --allow-extra-chr --make-bed --out GWAs_raw/snps_fil1

本研究包含三种表型信息，需要使用Plink插入原始数据：
添加表型（生态型）信息：

plink --bfile GWAs_raw/snps_fil1 --pheno POP.ecotypes --allow-extra-chr --make-bed --out GWAs_raw/snps_fil1_p

初始SNP数据不包含SNP ID，本文以染色体编号:位置信息作为染色体编号
添加SNP ID：

plink -bfile GWAs_raw/snps_fil1_p --set-missing-var-ids @:# --allow-extra-chr --make-bed --out GWAs_raw/RT_snps

2 关联分析

PLINK的--assoc参数是进行关联分析的参数，--adjust将分析的原始P值进行修正，由于研究设计的材料为植物，不涉及性别信息，同时缺少染色体编号（主要是格式不对，现在改比较占用资源，最后更改更方便），因此需要添加--allow-extra-chr和--allow-no-sex参数:

plink -bfile GWAs_raw/RT_snps --assoc --adjust --allow-extra-chr --allow-no-sex --out GWAs_raw/RT_snps

将结果中的染色体编号替换为纯数字：

# to substitute chrm ID
bash script/chr_tran.sh GWAs_raw/RT_snps.qassoc.adjusted GWAs_raw/RT_snps1.qassoc.adjusted

添加变异位点位置信息：

# add BP info
nohup bash script/add_BP.sh GWAs_raw/RT_snps1.qassoc.adjusted GWAs_raw/RT_snps2.qassoc.adjusted &

script:
chr_tran.sh:

INPUT=$1
OUT=$2
sed "s/Chrom1/1/g" $INPUT |awk '{print$0}' \
        |sed "s/Chrom2/2/g" |awk '{print$0}' \
        |sed "s/Chrom3/3/g" |awk '{print$0}' \
        |sed "s/Chrom4/4/g" |awk '{print$0}' \
        |sed "s/Chrom5/5/g" |awk '{print$0}' \
        |sed "s/Chrom6/6/g" |awk '{print$0}' \
        |sed "s/Chrom7/7/g" |awk '{print$0}' \
        |sed "s/Chrom8/8/g" |awk '{print$0}' \
        |sed "s/Chrom9/9/g" |awk '{print$0}' \
        |sed "s/Chrom10/10/g" |awk '{print$0}' \
        |sed "s/Chrom11/11/g" |awk '{print$0}' \
        |sed "s/Chrom12/12/g" |awk '{print$0}' > $OUT

add_BP.sh:

INPUT=$1
OUTPUT=$2
while read -r line; do
    fields=($line)
    if [[ ${fields[1]} == "SNP" ]]; then
        echo "$line BP"
    else
        second_column=${fields[1]}
        IFS=':' read -ra values <<< "$second_column"
        new_column_value=${values[1]}
        echo "$line $new_column_value"
    fi
done < $INPUT > $OUTPUT

结果可视化：

nohup Rscript --no-save QQ_man.R >> Manhattan.log&

install.packages("qqman",repos="https://mirrors.tuna.tsinghua.edu.cn/CRAN/",lib="~" ) 
library("qqman",lib.loc="~") 

results_as <- read.table("RT_snps2.qassoc.adjusted", head=TRUE)


##  BONF
#QQ
jpeg("RT_BONF.jpeg")
qq(results_as$BONF, main = "Q-Q plot of GWAS BONF : assoc")
dev.off()
#Manhattan 
jpeg("RT_man_BONF.jpeg")
manhattan(results_as,chr="CHR",bp="BP",p="BONF",snp="SNP", main = "Manhattan plot: assoc BONF")
dev.off()


##  GC
#QQ
jpeg("RT_GC.jpeg")
qq(results_as$GC, main = "Q-Q plot of GWAS GC : assoc")
dev.off()
#Manhattan 
jpeg("RT_man_GC.jpeg")
manhattan(results_as,chr="CHR",bp="BP",p="GC",snp="SNP", main = "Manhattan plot: assoc GC")
dev.off()


##  UNADJ
#QQ
jpeg("RT_UNADJ.jpeg")
qq(results_as$UNADJ, main = "Q-Q plot of GWAS UNADJ : assoc")
dev.off()
#Manhattan 
jpeg("RT_man_UNADJ.jpeg")
manhattan(results_as,chr="CHR",bp="BP",p="UNADJ",snp="SNP", main = "Manhattan plot: assoc UNADJ")
dev.off()


##  HOLM
#QQ
jpeg("RT_HOLM.jpeg")
qq(results_as$HOLM, main = "Q-Q plot of GWAS HOLM : assoc")
dev.off()
#Manhattan 
jpeg("RT_man_HOLM.jpeg")
manhattan(results_as,chr="CHR",bp="BP",p="HOLM",snp="SNP", main = "Manhattan plot: assoc HOLM")
dev.off()

##  FDR_BH
#QQ
jpeg("RT_FDR_BH.jpeg")
qq(results_as$FDR_BH, main = "Q-Q plot of GWAS FDR_BH : assoc")
dev.off()
#Manhattan 
jpeg("RT_man_FDR_BH.jpeg")
manhattan(results_as,chr="CHR",bp="BP",p="FDR_BH",snp="SNP", main = "Manhattan plot: assoc FDR_BH")
dev.off()


##  FDR_BY
#QQ
jpeg("RT_FDR_BY.jpeg")
qq(results_as$FDR_BY, main = "Q-Q plot of GWAS FDR_BY : assoc")
dev.off()
#Manhattan 
jpeg("RT_man_FDR_BY.jpeg")
manhattan(results_as,chr="CHR",bp="BP",p="FDR_BY",snp="SNP", main = "Manhattan plot: assoc FDR_BY")
dev.off()

3 另一种可视化方法

install.packages("CMplot",repos="https://mirrors.tuna.tsinghua.edu.cn/CRAN/",lib="~")
install.packages("qqman",repos="https://mirrors.tuna.tsinghua.edu.cn/CRAN/",lib="~")

library("CMplot")
library("qqman")

results_as <- read.table("Rcau_FDR_BH.assoc", head=TRUE)


CMplot(results_as,plot.type="q", threshold=0.05)

CMplot(results_as,plot.type="m",
	threshold=c(0.01,0.05)/nrow(results_as),
	col=c("grey30","#FFACAA",
		"grey30","grey60",
		"grey30","grey60",
		"grey30","grey60",
		"grey30","grey60",
		"grey30","grey60")
	threshold.col=c('red','black'), 
	threshold.lty=c(1,2),
	threshold.lwd=c(1,1),
	amplify=T, 
	signal.cex=c(1,1),
	signal.pch=c(20,20),
	signal.col=c("red","orange"))

结果：