制作一个函数来获取数据框R中多行的百分比

Question

我有一个包含数百个样本和数千个变量的数据框。但是这里我给出一个简单的数据框（my_data）作为说明。我想根据基因所属的集群（基因可以在多个集群中）获得变量相对于gene_count总和的百分比。我知道如何获取每个使用数据框操作的百分比。但是，由于我是编码新手，我正在尝试创建一个函数来获取百分比。谁能帮助如何使用我的函数（百分比）获得每个基因的百分比？结果将是“百分比”列上的百分比。非常感谢。

gene = c("CD63", "PTN", "MT2A", "PTGDS", "DBI", "TIMP1", "COX6C", "APLP2", "PTN", "GPC1")
gene_count = c(10, 15, 5, 15, 10, 25, 5, 5, 5, 5)
cluster = c(1,2, 3, 5, 7, 8, 9, 3, 6, 4 )
percent = c(0.1, 0.15, 0.5, 0.15, 0.1, 0.25, 0.05, 0.05, 0.05, 0.05)
my_data = data.frame(gene, gene_count, cluster, percent)
my_data


percent = function(gene, cluster){
for (gene in c(data$gene)){
if (data$gene == gene & data$cluster == cluster)
print(data$gene_count[which(data$gene == gene & data$cluster == cluster)]/sum(data$gene_count))
else print("Gene is not expressed in this cluster")
  }
}

Answer 1

这可能有用。

percent_in_total

percent_in_total = function(data, gene_in, cluster_in){
  data %>% 
  filter(gene == gene_in & cluster == cluster_in) %>% 
  .[["gene_count"]]/sum(gene_count)
}
percent_in_total(my_data,"PTN",2)
[1] 0.15

# data.table version
library(data.table)
percent_in_total = function(data, gene_in, cluster_in){
  setDT(data)[,
              .SD[gene == gene_in & cluster == cluster_in, gene_count]] / sum(gene_count)
}
percent_in_total(my_data,"PTN",2)
[1] 0.15

percent_in_cluster

我更喜欢使用data.table 的语法。解释这个过程，对于基因MT2A，.SD[gene == gene_in & cluster == cluster_in, gene_count]是gene_count = 5的编号，sum(.SD[cluster==cluster_in,gene_count])]是cluster = 3的总基因编号，即5+5。

library(data.table)
percent_in_cluster = function(data, gene_in, cluster_in){
  setDT(data)[,
              .SD[gene==gene_in & cluster==cluster_in, gene_count] / sum(.SD[cluster==cluster_in,gene_count])]
}
percent_in_cluster(my_data,"MT2A",3)
[1] 0.5

Answer 2

你可以这样写函数：

percent <- function(data, my_gene, my_cluster) {
  sub_data <- subset(data, gene == my_gene & cluster == my_cluster)
  if(nrow(sub_data)) sum(sub_data$gene_count)/sum(data$gene_count)
  else cat("Gene is not expressed in this cluster")
}

percent(my_data, 'PTN', 2)
#[1] 0.15

percent(my_data, 'ABC', 2)
#Gene is not expressed in this cluster