Mapping SNP coordinates to gene coordinates is too slow - perl

Question

I have two tab-delimited files like these
File 1 (these are Single Nucleotide Polymorphism (SNP) positions)
Chr1 26690
Chr1 33667
Chr1 75049
.
.
Chr2 12342
Chr2 32642
Chr2 424421
.
.
File 2 (these are gene start and end coordinates)
Chr1 2903 10817 LOC_Os01g01010
Chr1 2984 10562 LOC_Os01g01010
Chr1 11218 12435 LOC_Os01g01019
Chr1 12648 15915 LOC_Os01g01030
Chr1 16292 18304 LOC_Os01g01040
Chr1 16292 20323 LOC_Os01g01040
Chr1 16321 20323 LOC_Os01g01040
Chr1 16321 20323 LOC_Os01g01040
Chr1 22841 26971 LOC_Os01g01050
Chr1 22841 26971 LOC_Os01g01050
.
.
What I want is to match SNPs in file 1 to genes in file 2. The script should match the string in the first column of the files, and if they match it should then find which gene in the file 2 contains the corresponding SNP and return the locus ID from the fourth column of File 2.
Here's the script I have written
use strict;
my $i1 = $ARGV[0]; # SNP
my $i2 = $ARGV[1]; # gene coordinate
open(I1, $i1);
open(I2, $i2);
my #snp = ();
my #coor = ();
while( <I1> ) {
push(#snp, $_);
}
while ( <I2> ) {
push(#coor, $_);
}
for ( my $i = 0; $i <= $#snp; $i++ ) {
my #snp_line = split "\t", $snp[$i];
for ( my $j = 0; $j <= $#coor; $j++ ) {
my #coor_line = split "\t", $coor[$i];
if ( $snp_line[0] eq $coor_line[0] ) {
if ( $snp_line[1] >= $coor_line[1] && $snp_line[1] <= $coor_line[2] ) {
print "$snp_line[0]\t$snp_line[1]\t$coor_line[3]\n";
goto a;
}
}
}
a:
}
The problem is that obviously this is not the best way to do it as it iterates over all the ~60,000 lines in file 2 for each SNP in line 1. Also, it ran overnight and did not go past Chr1; we have upto Chr12.

Answer 1

You could work with these files when reformatted as UCSC BED format, using a toolkit like BEDOPS that does efficient set operations on sorted BED files.
Convert your first file of SNPs to a sorted BED file:
$ awk -v OFS="\t" '{ print $1, $2, ($2+1); }' snps.txt | sort-bed - > snps.bed
Sort the genes ("file 2"):
$ sort-bed genes.unsorted.txt > genes.bed
Map SNPs to genes:
$ bedmap --echo --echo-map-id-uniq --delim '\t' snps.bed genes.bed > answer.bed
If you need to, you can strip the end position of the SNP from the answer:
$ cut -f1,2,4 answer.bed > answer.txt
These tools will run very fast, usually within a few moments.
I would not use Perl or Python to do these kinds of set operations, unless I was doing some kind of academic exercise.

Answer 2

Here is a working script, the one posted above had bugs
use strict;
my $i1=$ARGV[0]; # SNP
my $i2=$ARGV[1]; # gene coordinate
open(I1,$i1);
open(I2,$i2);
my #snp=();
my #coor=();
while(<I1>)
{
push(#snp,$_);
}
while(<I2>)
{
push(#coor,$_);
}
for(my $i=0;$i<=$#snp;$i++)
{
my #snp_line = split "\t",$snp[$i];
for(my $j=0;$j<=$#coor;$j++)
{
my #coor_line = split "\t",$coor[$j];
if ($snp_line[0] eq $coor_line[0])
{
if ($snp_line[1] >= $coor_line[1] && $snp_line[1] <= $coor_line[2])
{
print "$snp_line[0]\t$snp_line[1]\t$coor_line[3]\n";
}
}
}
}
This one does the job.