################################################################################# ###Data Load and Overview######################################################## ################################################################################# getwd() #현재의 작업공간 확인하기 setwd("C:/R_data/lecture10") #제 10강 데이터가 있는 곳으로 이동 dir() ################################################################################# ################################################################################# ###각각의 파일을 변수로 불러오기### cli <- read.csv("JAMA_Clinical.csv", header=T, na.strings="null") mut <- read.delim("JAMA_Mutation.txt", header=T) met <- read.delim("JAMA_BRCA1_Methyl.txt",header=T) exp <- read.delim("JAMA_Expression.txt", header=T) ################################################################################## ################################################################################## ###데이터의 구조와 크기 파악하기### ls() #loading한 데이터 확인 head(cli) #first 1~6 row 만 보기 str(cli) #구조확인 dim(cli) #dimension 확인 summary(cli) #basic statistics 확인 head(mut) head(met) head(exp) ################################################################################# ################################################################################# ####BRCA1/2 mutation을 가진 sample 추출하기### brca1.mut <- unique(as.character(mut[which(mut[,2] =="BRCA1" ),1])) #mut table에서 gene 이름이 brca1인 sample 추출 brca2.mut <- unique(as.character(mut[which(mut[,2] =="BRCA2" ),1])) brca1.mut #출력해서 확인 brca2.mut ################################################################################# ################################################################################# ###BRCA1 expression level 추출하기### brca1.exp <- as.numeric(exp[which(exp[,1]=="BRCA1"),-1]) # BRCA1 gene에 대한 expression level 추출 head(brca1.exp) ###BRCA1 expression과 상관관계가 높은 methylation probe 추출하기### apply(met[,-1], 1, function(x)cor(x, brca1.exp)) exp.cor <- apply(met[,-1], 1, function(x)cor(x, brca1.exp)) #apply 함수를 이용하여 BRCA1 expression level과 각각의 methylation probe의 beta value간의 pearson correlation coefficient를 구한다. TestMatrix <- met[which(abs(exp.cor) > 0.3),-1] #BRCA1 expression이 달라진 methylation prove를 찾는다. heatmap(as.matrix(TestMatrix)) #methylation pattern이 다른 샘플을 찾기 위해heatmap을 그린다. ############################################################################### ############################################################################### ###cluster dendrogram을 이용해 BRCA1 methylation 이 달라진 샘플 찾기### AA<- hclust(dist(t(TestMatrix))) plot(AA) M<-identify(AA) #R graphic 창에 나타난 dendrogram에서 가장 왼쪽에 차이나는 샘플들을 선택(왼쪽클릭)한 후 stop(오른쪽클릭해서 나타나는 stop선택) M #확인 brca1.met<- names(met[,as.numeric(M[[1]])]) #BRCA1 methylation 이 다른 sample을 변수에 할당 brca1.met #확인 ############################################################################### ############################################################################### ###survival analysis 하기### type <- c(rep('wild', 317)) #먼저 전체 type이 wild 인 변수를 만든다. names(type)<-cli[,1] #column 명을 환자 id 로 바꿔준다. brca1.met<-gsub(".","-",brca1.met, fixed=T) #샘플의 이름 표기가 서로 다르므로 이를 바꿔준다. brca1.mut #분석할 테이블 확인 brca2.mut brca1.met #교재 오타 type[brca1.mut] #샘플별 type을 바꿔줌 type[brca1.met] = 'brca1.met' type[brca1.mut] = 'brca1.mut' type[brca2.mut] = 'brca2.mut' library(survival) #survival 분석을 위한 함수를 불러온다. out <- survfit(Surv(cli$days_to_last_followup) ~ type) color <- c("green","blue", "red","black") #4개 군을 나타낼 색깔을 정해둔다. plot(out, col=color , main="Association of BRCA1/2 Mutations with Survival", xlab="Time,days", ylab="Proportion") #plot legend(4200, 0.9, c("BRCA1 mutation","BRCA1 methylation","BRCA2 mutation","wild"), col=color, lty=1, lwd=3) ###############################################################################