create data.fame satisfying some condition - r

Question

Input
final_table =
Chr start end num seg.mean seg.mean.1 seg.mean.2
1 68580000 68640000 A8430 0.7000 0.1440 0.1032
1 115900000 116260000 B8430 0.0039 2.7202 2.7202
1 173500000 173680000 C5 -1.7738 -0.0746 -0.2722
How can I make a new data.frame where the values of columns 5 through 7 are set to:
-1, if value < -0.679
0, if -0.679 <= value <= 0.450
+1, if value > 0.450
Expected output
Chr start end num seg.mean seg.mean.1 seg.mean.2
1 68580000 68640000 A8430 1 0 0
1 115900000 116260000 B8430 0 1 1
1 173500000 173680000 C5 -1 0 0

Answer 1

try this:
# read the data in
df <- read.table(header = TRUE, text="Chr start end num seg.mean seg.mean.1 seg.mean.2
1 68580000 68640000 A8430 0.7000 0.1440 0.1032
1 115900000 116260000 B8430 0.0039 2.7202 2.7202
1 173500000 173680000 C5 -1.7738 -0.0746 -0.2722")
# get the column-names of the columns you wanna change
cols <- names(df[5:length(df)])
# set a function for the different values you want for the value-ranges
fun_cond <- function(x) {
ifelse(x < -0.679 , -1, ifelse(
x >= -0.679 & x <= 0.450, 0, 1))
}
# copy the data-frame so the old one doesnt get overwritten
new_df <- df
# work with data-table to apply the function to the columns
library(data.table)
setDT(new_df)[ , (cols) := lapply(.SD, fun_cond), .SDcols = cols]
output:
Chr start end num seg.mean seg.mean.1 seg.mean.2
1: 1 68580000 68640000 A8430 1 0 0
2: 1 115900000 116260000 B8430 0 1 1
3: 1 173500000 173680000 C5 -1 0 0
same thing without using any additional packages:
cols <- names(df[5:length(df)])
fun_cond <- function(x) {
ifelse(x < -0.679 , -1, ifelse(
x >= -0.679 & x <= 0.450, 0, 1))
}
new_df <- df
new_df[5:length(df)] <- lapply(new_df[5:length(df)], fun_cond)

Answer 2

I'd use the cut function and apply it to the last three columns individually.
Here's a simple example:
original = data.frame(a=c(rep("A", 2), rep("B", 2)), seg.mean=c(-1, 0, 0.4, 0.5));
original$segmented = cut(original$seg.mean, c(-Inf, -0.679, 0.450, Inf), labels = c(-1,0,1))
One thing to be careful about: the new column will be a factor. If you need numerical values, you may need to apply as.numeric to it.
You can also try to use labels=FALSE which will give you numerical values (but likely 1,2,3 rather than -1,0,1). You can fix that by subtracting 2:
original$segmented = cut(original$seg.mean, c(-Inf, -0.679, 0.450, Inf), labels = FALSE)-2

Answer 3

You can directly replace fields in the data frame by subsetting
df[, 5:7] <- ifelse(df[, 5:7] < -0.679, -1,
ifelse(df[, 5:7] < 0.450, 0,
1))