MR循环——忽略502报错

添加了去除混淆SNP和筛选F值操作
另外输出的文件mr_KEEP=TRUE

library('TwoSampleMR')
id_list = read.table('C:/Users/DELL/Desktop/cs/邱师兄越高/expo_list.txt',header = T,sep = '\t')
remove_snp = read.table('./remove_snp.txt', header = T, sep = '\t')


# 设置最大重试次数和每次重试等待的时间
max_retries <- 55
wait_time <- 5

# 定义一个函数用来连接服务器
connect_to_server <- function() {
  # 尝试连接服务器
  connection <- try({extract_instruments(outcomes = i)}, silent = TRUE)
  
  # 如果连接失败，则重试
  retries <- 1
  while (inherits(connection, "try-error") && retries <= max_retries) {
    # 等待一段时间后重新连接
    Sys.sleep(wait_time)
    connection <- try({extract_instruments(outcomes = i)}, silent = TRUE)
    retries <- retries + 1
  }
  
  # 返回连接结果
  return(connection)
}

# 循环遍历所有的 exposure_id
for (i in id_list$exposure_id) {
  # 连接服务器
  exposure <- connect_to_server()
  try(dev.off(), silent = TRUE)
  
  # 检查连接是否成功
  if (inherits(exposure, "try-error")) {
    # 如果连接失败，则输出错误信息并继续下一个循环
    cat("Unable to connect to server for exposure", i, "\n")
    try(dev.off(), silent = TRUE)
    next
  }
  print(exposure$exposure[1])
  # 提取 outcome 数据
  exposure$RR <- 2 * exposure$eaf.exposure * (1 - exposure$eaf.exposure) * ((exposure$beta.exposure)^2)
  exposure$F <- (exposure$beta.exposure)^2 / (exposure$se.exposure)^2
  exposure <- exposure[exposure$F > 10, ]
  print((table(exposure$id.exposure)))
  print(format(sum(exposure$RR), scientific = TRUE, digits = 2))
  
  outcome <- extract_outcome_data(snps = exposure$SNP, outcomes = 'ukb-b-19813')
  
  # 检查 outcome 数据是否有效
  if (is.null(outcome)) {
    # 如果 outcome 数据无效，则输出错误信息并继续下一个循环
    cat("Unable to extract outcome data for exposure", i, "\n")
    try(dev.off(), silent = TRUE)
    next
  }
  
  # 继续执行后续分析
  dat <- harmonise_data(exposure_dat = exposure, outcome_dat = outcome)
  dat$F <- (dat$beta.exposure)^2 / (dat$se.exposure)^2
  dat <- dat[dat$mr_keep==TRUE,]
  dat <- dat[!(dat$SNP %in% remove_snp$SNP),]
  print(table(dat$mr_keep))
  # filename <- paste0("./SNP/dat", i, '.txt')
  # write.table(dat,file = filename, quote = F, row.names = F, sep = '\t')
  # a <- generate_odds_ratios(mr_res = mr(dat))
  # b <- mr_heterogeneity(dat)
  # c <- mr_pleiotropy_test(dat)
  # filenamea <- paste0('./mr/mr_', i, '.txt')
  # filenameb <- paste0('./heter/heter_', i, '.txt')
  # filenamec <- paste0('./pleio/pleio_', i, '.txt')
  # write.table(a, file = filenamea, , quote = F, sep = '\t')
  # write.table(b, file = filenameb, quote = F, sep = '\t')
  # write.table(c, file = filenamec, quote = F, sep = '\t')
  # 
  # filename <- paste0("./fig/outfig_", i, '.pdf')
  # try(dev.off(), silent = TRUE)
  # pdf(file = filename, onefile = TRUE)
  # a <- mr_scatter_plot(mr_results = mr(dat, method_list = c("mr_ivw", "mr_egger_regression", "mr_weighted_median")), dat)
  # b <- mr_funnel_plot(singlesnp_results = mr_singlesnp(dat))
  # c <- mr_leaveoneout_plot(leaveoneout_results = mr_leaveoneout(dat))
  # print(a)
  # print(b)
  # print(c)
  # dev.off()

  try(dev.off(), silent = TRUE)
  Sys.sleep(wait_time)
  print('over')
}