在R语言中实现两个列表对应列元素相乘的方法

问题背景 #

给定两个R列表list1和list2：

• list1的每个元素是仅含1行的tibble，列名为SNP标识，值为系数
• list2的每个元素是多行tibble，包含样本信息列（name、ENSG、expr）和对应SNP列，值为样本的基因型计数
• 两个列表的元素数量一致，对应元素的SNP列名完全匹配

需求：将list1中每个元素的各SNP系数，与list2对应元素的同列SNP值逐一相乘，保留list2的样本信息列，替换SNP列为相乘后的结果，同时保留数据结构。

输入数据 #

list1 #

list1 <- list(structure(list(chr22_20230714_G_A_b38 = 0.0000953181301665087, 
    chr22_20230737_G_A_b38 = -0.00124036704551427, chr22_20231229_T_A_b38 = 0.000808061558738542, 
    chr22_20231474_G_A_b38 = 0.000387528601423933, chr22_20231667_C_G_b38 = -0.000120624028990859), row.names = c(NA, 
-1L), class = c("tbl_df", "tbl", "data.frame")), structure(list(
    chr22_47157062_G_A_b38 = 0.00000909931572319958, chr22_47157212_G_A_b38 = -0.000124084106569373, 
    chr22_47157394_C_G_b38 = -0.0000752774417069946, chr22_47157559_G_A_b38 = 0.0000808446315377557, 
    chr22_47157607_T_C_b38 = 0.000237979025556899), row.names = c(NA, 
-1L), class = c("tbl_df", "tbl", "data.frame")))

list2 #

list2 <- list(structure(list(name = c("HG00096", "HG00097", "HG00099", 
"HG00100", "HG00101"), ENSG = c("ENSG00000040608", "ENSG00000040608", 
"ENSG00000040608", "ENSG00000040608", "ENSG00000040608"), expr = c(-0.5186894, 
0.6170779, -0.5786774, 0.07324268, -0.7579184), chr22_20230714_G_A_b38 = c(1L, 
1L, 1L, 2L, 1L), chr22_20230737_G_A_b38 = c(0L, 0L, 0L, 0L, 0L
), chr22_20231229_T_A_b38 = c(1L, 0L, 1L, 0L, 1L), chr22_20231474_G_A_b38 = c(0L, 
1L, 0L, 0L, 0L), chr22_20231667_C_G_b38 = c(1L, 1L, 1L, 2L, 1L
)), row.names = c(NA, -5L), class = c("tbl_df", "tbl", "data.frame"
)), structure(list(name = c("HG00096", "HG00097", "HG00099", 
"HG00100", "HG00101"), ENSG = c("ENSG00000054611", "ENSG00000054611", 
"ENSG00000054611", "ENSG00000054611", "ENSG00000054611"), expr = c(-0.5555929, 
0.1600335, 0.4027508, -0.6028474, 2.271097), chr22_47157062_G_A_b38 = c(0L, 
1L, 0L, 0L, 0L), chr22_47157212_G_A_b38 = c(0L, 0L, 1L, 1L, 2L
), chr22_47157394_C_G_b38 = c(0L, 1L, 1L, 1L, 2L), chr22_47157559_G_A_b38 = c(0L, 
1L, 0L, 0L, 0L), chr22_47157607_T_C_b38 = c(0L, 1L, 1L, 1L, 2L
)), row.names = c(NA, -5L), class = c("tbl_df", "tbl", "data.frame"
)))

解决方案代码 #

使用purrr包的map2函数批量处理两个列表的对应元素，结合dplyr进行列运算：

library(purrr)
library(dplyr)

out <- map2(list2, list1, function(df2, df1) {
  # 获取SNP列名
  snp_cols <- names(df1)
  # 对每个SNP列执行相乘运算，同时保留样本信息列
  df2 %>%
    mutate(across(all_of(snp_cols), ~ .x * df1[[cur_column()]])) %>%
    # 匹配示例精度保留小数位数（可选）
    mutate(across(all_of(snp_cols), ~ round(.x, 6)))
})

验证结果 #

运行代码后，out的结构和值与预期输出一致：

print(out)

输出结果：

out <- list(structure(list(name = c("HG00096", "HG00097", "HG00099", 
"HG00100", "HG00101"), ENSG = c("ENSG00000040608", "ENSG00000040608", 
"ENSG00000040608", "ENSG00000040608", "ENSG00000040608"), expr = c(-0.5186894, 
0.6170779, -0.5786774, 0.07324268, -0.7579184), chr22_20230714_G_A_b38 = c(0.000095, 
0.000095, 0.000095, 0.000191, 0.000095), chr22_20230737_G_A_b38 = c(0, 
0, 0, 0, 0), chr22_20231229_T_A_b38 = c(0.000808, 0, 0.000808, 
0, 0.000808), chr22_20231474_G_A_b38 = c(0, 0.000388, 0, 0, 0
), chr22_20231667_C_G_b38 = c(-0.000121, -0.000121, -0.000121, 
-0.000241, -0.000121)), row.names = c(NA, -5L), class = c("tbl_df", 
"tbl", "data.frame")), structure(list(name = c("HG00096", "HG00097", 
"HG00099", "HG00100", "HG00101"), ENSG = c("ENSG00000054611", 
"ENSG00000054611", "ENSG00000054611", "ENSG00000054611", "ENSG00000054611"
), expr = c(-0.5555929, 0.1600335, 0.4027508, -0.6028474, 2.271097
), chr22_47157062_G_A_b38 = c(0, 0.000009, 0, 0, 0), chr22_47157212_G_A_b38 = c(0, 
0, -0.000124, -0.000124, -0.000248), chr22_47157394_C_G_b38 = c(0, 
-0.000075, -0.000075, -0.000075, -0.000151), chr22_47157559_G_A_b38 = c(0, 
0.000081, 0, 0, 0), chr22_47157607_T_C_b38 = c(0, 0.000238, 
0.000238, 0.000238, 0.000476)), row.names = c(NA, -5L), class = c("tbl_df", 
"tbl", "data.frame")))

内容的提问来源于stack exchange，提问作者Maya_Cent