在R中应用t.test数据帧

Question

I am new to programming, and have little idea how to approach this, any help would be much appreciated.

I must use a two-sample t-test to compare two sets of data, c and t , each of which is divided into 6 sub columns in the same dataframe, in Excel the data appears something like this:

name|c1|c2|c3|c4|c5|c6|t1|t2|t3|t4|t5|t6

"name" is the same for each row, but differs between rows. The columns c1-t6 contain numeric values, which differ between each row and column.

Each row must be tested individually, comparing the c subgroup to the t subgroup.

How would I go about doing this? I'm guessing a loop will be needed?

Answer 1

Using @thelatemail's input, you would most likely do the following, illustrated here with a reproducible example. df is your data.frame and as I work with dplyr , I'll use this here too.

require(dplyr)
df <- data.frame(
+     name = sample(letters[1:10]),
+     c1 = sample(1:10),
+     c2 = sample(1:10),
+     t1 = sample(1:10), 
+     t2 = sample(1:10))
df
   name c1 c2 t1 t2
1     i  7  3  8  2
2     h  6  4  4  8
3     g  4  6  6  5
4     b  5  1  9 10
5     a  9  5  3  7
6     j  8  9  5  3
7     d 10  8 10  4
8     c  2  2  2  1
9     e  1 10  7  6
10    f  3  7  1  9
df1 <- df %>% select(contains("c"))
df2 <- df %>% select(contains("t"))
Map(t.test, as.data.frame(df1), as.data.frame(df2))

But, I'm not entirley sure that's what you want to do, as this seems to loop the function over columns and not rows. Thus, a bit of a hacky solution (please someone show me an easier way), I would do the following:

require(tidyr)
df2 <- gather(df, condition, measurement, c1:t2)
df3 <- spread(df2, name, measurement)
df3$condition2 <- ifelse(grepl("c", df3$condition), "c", "t")
#check dimensions of new df3
for(i in 2:11){cat(colnames(df3)[i],'\n')
+                y <- df3[, i]
+                res <- t.test(y~df3$condition2, var.equal=T)
+                print(res)
+ }

note: I've added the var.equal=T assuming you want to do a two sample t.test()

I believe this gives you the t.test for your data you desire.

Answer 2

Assuming the unpaired two group t-test , consider using the mapply function, the multivariate version of sapply which applies FUN to the first elements of each argument, the second elements, the third elements, and so on.

# DF SPLIT BETWEEN EACH CONTROL AND TREATMENT
controls <- df[c(grep("c", names(df)))]        # ALL C COLS
treatments <- df[c(grep("t", names(df)))]      # ALL T COLS

# MAPPLY USING TTEST
tstats_m <- mapply(ttest, var1=controls, var2=treatments)
tstats_m <- as.data.frame(tstats_m)

# MAPPLY USING DEFINED FUNCTION TTEST
tfunc <- function(var1, var2){
            t.test(var1, var2)            
          }
tstats_m <- mapply(tfunc, var1=controls, var2=treatments)

Alternatively, below is the traditional for loop that iterates results of each test:

for (i in 1:ncol(controls){
  print(paste0("Two-sample t-test c", i, " = t", i))
  print(t.test(controls[paste0("c", i)], treatments[paste0("t", i)]))
}