grep "[ATGC]>[ATGC]" file.txt > output.txt

using simple grep command you can extract the line based on pattern. for e.g

grep "C>T" file.txt >output.txt

and grep can be done in R as well

a <- c("c.2455C>T","c.2455C>G","c.2455C>A")
b <- grep(pattern = "[C]>[TG]",a,perl = TRUE,value = TRUE) # as  Pierre suggested
b
[1] "c.2455C>T" "c.2455C>G"

Assuming you want just the mutation type part, you can strip the string with sed -E 's|.+[0-9_]+||g'

$ echo -e "c.2455C>T\nc.234A>G\nc.6345TTT>G\nc.4375_4376insACCT\nc.4375_4379del"
c.2455C>T
c.234A>G
c.6345TTT>G
c.4375_4376insACCT
c.4375_4379del
$ echo -e "c.2455C>T\nc.234A>G\nc.6345TTT>G\nc.4375_4376insACCT\nc.4375_4379del" | sed -E 's|.+[0-9_]+||g'
C>T
A>G
TTT>G
insACCT
del

Note: If your input is any HGVS cDNA mutation, you may get del,ins,delins etc. and not just substitutions (e.g. A>T).

In R just use gsub

> gsub("c.4375_4376insACCT", pattern=".+[0-9_]+", replacement="")
[1] "insACCT"