merge partial matched strings - r

Question

I am struggling with trying to combine partially matched strings from two files.
File 1 contains a list of unique strings. These strings are partially matched to a number of strings in File 2. How do I merge the rows in file 1 with file 2 for every matched case
File1
mmu-miR-677-5p_MIMAT0017239
mmu-miR-181a-1-3p_MIMAT0000660
File2
mmu-miR-677-5p_CTTCAGTGATGATTAGCTTCTGA
mmu-miR-677-5p_CTTCAGTGATGATTAGCTTCTGACT
mmu-miR-677-5p_TTCAGTGATGATTAGCTTCTGACT
mmu-miR-181a-1-3p_ACCATCGACCGTTGATTGTAC
mmu-miR-181a-1-3p_ACCATCGACCGTTGATTGTACC
Desired output
mmu-miR-677-5p_MIMAT0017239 mmu-miR-677-5p_CTTCAGTGATGATTAGCTTCTGA
mmu-miR-677-5p_MIMAT0017239 mmu-miR-677-5p_CTTCAGTGATGATTAGCTTCTGACT
mmu-miR-677-5p_MIMAT0017239 mmu-miR-677-5p_TTCAGTGATGATTAGCTTCTGACT
mmu-miR-181a-1-3p_MIMAT0000660 mmu-miR-181a-1-3p_ACCATCGACCGTTGATTGTAC
mmu-miR-181a-1-3p_MIMAT0000660 mmu-miR-181a-1-3p_ACCATCGACCGTTGATTGTACC
I have tried using pmatch() in R, but don't get it right. I looks like something perl would handle??
Maybe something like this:
perl -ne'exec q;perl;, "-ne", q $print (/\Q$.$1.q;/?"$. YES":$. .q\; NO\;);, "file2" if m;^(.*)_pat1;' file1

Answer 1

This is a brief Perl solution, which saves all the data from file1 in a hash and then retrieves it as file2 is scanned
use strict;
use warnings;
use autodie;
my #files = qw/ file1.txt file2.txt /;
my %file1 = do {
open my $fh, '<', $files[0];
map /([^_]+)_(\S+)/, <$fh>;
};
open my $fh, '<', $files[1];
while (<$fh>) {
my ($key) = /([^_]+)/;
printf "%-32s%s", "${key}_$file1{$key}", $_;
}
output
mmu-miR-677-5p_MIMAT0017239 mmu-miR-677-5p_CTTCAGTGATGATTAGCTTCTGA
mmu-miR-677-5p_MIMAT0017239 mmu-miR-677-5p_CTTCAGTGATGATTAGCTTCTGACT
mmu-miR-677-5p_MIMAT0017239 mmu-miR-677-5p_TTCAGTGATGATTAGCTTCTGACT
mmu-miR-181a-1-3p_MIMAT0000660 mmu-miR-181a-1-3p_ACCATCGACCGTTGATTGTAC
mmu-miR-181a-1-3p_MIMAT0000660 mmu-miR-181a-1-3p_ACCATCGACCGTTGATTGTACC

Answer 2

Of course you may do it in R. Indeed, pmatching whole strings won't give you the desired result - you've got to match appropriate substrings.
I assume that in file 1 the first identifier is 677 and not 667, otherwise it's hard to guess the matching scheme (I assume your example is only a part of a bigger database).
file1 <- readLines(textConnection('mmu-miR-677-5p_MIMAT0017239
mmu-miR-181a-1-3p_MIMAT0000660'))
file2 <- readLines(textConnection('mmu-miR-677-5p_CTTCAGTGATGATTAGCTTCTGA
mmu-miR-677-5p_CTTCAGTGATGATTAGCTTCTGACT
mmu-miR-677-5p_TTCAGTGATGATTAGCTTCTGACT
mmu-miR-181a-1-3p_ACCATCGACCGTTGATTGTAC
mmu-miR-181a-1-3p_ACCATCGACCGTTGATTGTACC'))
library(stringi)
file1_id <- stri_extract_first_regex(file1, "^.*?(?=_)")
file2_id <- stri_extract_first_regex(file2, "^.*?(?=_)")
cbind(file1=file1[match(file2_id, file1_id)], file2=file2)
## file1 file2
## [1,] "mmu-miR-677-5p_MIMAT0017239" "mmu-miR-677-5p_CTTCAGTGATGATTAGCTTCTGA"
## [2,] "mmu-miR-677-5p_MIMAT0017239" "mmu-miR-677-5p_CTTCAGTGATGATTAGCTTCTGACT"
## [3,] "mmu-miR-677-5p_MIMAT0017239" "mmu-miR-677-5p_TTCAGTGATGATTAGCTTCTGACT"
## [4,] "mmu-miR-181a-1-3p_MIMAT0000660" "mmu-miR-181a-1-3p_ACCATCGACCGTTGATTGTAC"
## [5,] "mmu-miR-181a-1-3p_MIMAT0000660" "mmu-miR-181a-1-3p_ACCATCGACCGTTGATTGTACC"

Answer 3

You can agrep for fuzzy search. You should play with distance. Here I am fixing it manually to 11.
Basically I am doing this to extract lines number that matches each word in file1:
sapply(file1,agrep,file2,max=11)
$`mmu-miR-677-5p_MIMAT0017239`
[1] 1 2 3
$`mmu-miR-181a-1-3p_MIMAT0000660`
[1] 4 5
To get the result of a data.frame:
do.call(rbind,
lapply(file1,
function(x)
data.frame(file1=x,
file2=agrep(x,file2,max=11,value=T))))
file1 file2
1 mmu-miR-677-5p_MIMAT0017239 mmu-miR-677-5p_CTTCAGTGATGATTAGCTTCTGA
2 mmu-miR-677-5p_MIMAT0017239 mmu-miR-677-5p_CTTCAGTGATGATTAGCTTCTGACT
3 mmu-miR-677-5p_MIMAT0017239 mmu-miR-677-5p_TTCAGTGATGATTAGCTTCTGACT
4 mmu-miR-181a-1-3p_MIMAT0000660 mmu-miR-181a-1-3p_ACCATCGACCGTTGATTGTAC
5 mmu-miR-181a-1-3p_MIMAT0000660 mmu-miR-181a-1-3p_ACCATCGACCGTTGATTGTACC