R语言使用函数随机抽取并求均值和做T检验，最后输出随机抽取50次均值和pvalue的结果

1.输入数据：“5utr-5d做ABD中有RG4和没有RG4的TE之间的T检验.csv”
在这里插入图片描述

2.代码：

setwd("E:\\R\\Rscripts\\5UTR_extended_TE")
# 载入必要的库
library(tidyverse)
library(dplyr)
library(openxlsx)
# 读取数据
data <- read.csv("5utr-15d做ABD中有RG4和没有RG4的TE之间的T检验.csv", na.strings = "#N/A")

# 将所有的NA值转换为0
data <- data %>% mutate_all(~ifelse(is.na(.), 0, .))

############################################################  
# 调整后的process_scores函数1,适用于le1的个数小于ge1的个数且ave-le1大于ave-ge1的情况
############################################################
process_scores <- function(df, score_name, TE_name) {
  successful_seeds <- list() # 初始化一个列表来保存成功的seed值
  combined_samples_list <- list() # 新增：初始化一个列表来保存符合条件的组合数据框
  
  for (seed_val in 1:50) {
    set.seed(seed_val)
    ge1 <- df %>% filter(!!sym(score_name) >= 1) %>% select(!!sym(TE_name)) %>% mutate(Source = "ge1")
    le1 <- df %>% filter(!!sym(score_name) < 1) %>% select(!!sym(TE_name)) %>% mutate(Source = "sample_le1")
    
    sample_le1 <- sample_n(le1, nrow(ge1)) # 取单一样本进行比较
    
    t_test <- t.test(ge1[[1]], sample_le1[[1]])
    mean1 <- mean(ge1[[1]])
    mean2 <- mean(sample_le1[[1]])
    
    if (mean2 != mean1 && t_test$p.value <= 1) {
      successful_seeds[[paste0(seed_val, "_", score_name)]] <- list(
        seed = seed_val,
        mean1 = mean1,
        mean2 = mean2,
        pvalue = t_test$p.value
      )
      # 新增：将符合条件的ge1和sample_le1合并到一个数据框中，并保存到列表中
      combined_samples <- bind_rows(ge1, sample_le1)
      combined_samples_list[[paste0(seed_val, "_", score_name)]] <- combined_samples
    }
  }
  
  # 将成功的seeds信息转换为数据框
  if (length(successful_seeds) > 0) {
    successful_seeds_df <- bind_rows(successful_seeds, .id = "seed_score") %>% mutate(Comparison = seed_score)
  } else {
    successful_seeds_df <- tibble(Comparison = character(), mean1 = numeric(), mean2 = numeric(), pvalue = numeric())
  }
  
  # 新增：将combined_samples_list中的数据框合并或以其他形式输出
  combined_samples_output <- if (length(combined_samples_list) > 0) {
    # 例如，这里我们简单地将所有符合条件的数据框合并
    bind_rows(combined_samples_list)
  } else {
    # 如果没有符合条件的，则返回空数据框
    tibble()
  }
  
  return(list(successful_seeds = successful_seeds_df, combined_samples = combined_samples_output))
}

# 对AScore5d进行处理示例
results_AScore5d <- process_scores(data, "AScore5d", "ATe5d")
results_BScore5d <- process_scores(data, "BScore5d", "BTe5d")
results_DScore5d <- process_scores(data, "DScore5d", "DTe5d")

results_AScore5d_successful_seeds <- as.data.frame(results_AScore5d$successful_seeds)
results_BScore5d_successful_seeds <- as.data.frame(results_BScore5d$successful_seeds)
results_DScore5d_successful_seeds <- as.data.frame(results_DScore5d$successful_seeds)
# 打印出符合条件的successful_seeds结果进行检查
bind_results_AScore5d_successful_seeds<-cbind(results_AScore5d_successful_seeds,results_BScore5d_successful_seeds,results_DScore5d_successful_seeds)
write.xlsx(bind_results_AScore5d_successful_seeds, file = "5utr_bind_results_ABDScore15d_successful_seeds.xlsx")