When reading "File1.txt" you are not retaining the location information for each record. You could use a hash of hashes structure to do that. In the code below you can see this and I have added Data::Dumper output so that you can visualise the structure.

There seems to be some confusion in your code as to how to find the position for replacement. It would appear that you read the postition and new location from "File1.txt" but you then try to find things using rindex and then record the $current_base even though you don't seem to use it and are replacing it anyway. My code below just does a replacement based on the information in "File1.txt".

You process the records one at a time so I'm not sure why you use the %sequences hash. Why not just use a simple scalar to hold the record? I am assuming that the format you have posted for "File2.txt" is accurate with the data being broken into lines of 70 characters. Finding the correct position for the replacement needs to allow for the line breaks.

The code below inlines the input files using heredocs and the output file by opening a filehandle on a scalar reference. This is just to keep everything in the script but you can simply relace those constructs with actual files. I have changed the locations to asterisks just so they stand out in the output file so you can see the replacement happening. I have added a record where the replacement is not on the first line of data and another record where no change takes place.

use strict;
use warnings;

use Data::Dumper;

open my $in1FH, q{<}, \ <<END_File1 or die qq{open: < "File1.txt": $!\
+n};
155369268 5 * 
169212695 7 *
175387629 127 *
END_File1

my %positions = ();
while( <$in1FH> )
{
    my( $key, $posn, $locn ) = split;
    $positions{ $key } = {
       position => $posn,
       location => $locn,
       };
}
close $in1FH or die qq{close: < "File1.txt": $!\n};

print Data::Dumper->Dumpxs( [ \ %positions ], [ qw{ *positions } ] );

open my $in2FH, q{<}, \ <<END_File2 or die qq{open: < "File2.txt": $!\
+n};
>gi|155369268|ref|NM_001100917.1| After all
AAACAATGTCGATTCTATGATGCGAACGCAGCATTTCAGGGACTGGATGAGGAGCTTACGGTTTTTTACT
ACAGAATCATCAATATCTTGGAAGAAAAAGAATGTTAAGAAATAACAAAACAATAATTATTAAGTACTTT
>gi|169212695|ref|XM_001716884.1| Its the code 
AGACAAGCTTGTCCTGATGTTCCTTGCCCTGGCAGATGTTCAGGACCTTCCTTTGATTCAACCCCTCCAC
CTAAATGGCCCAAGCTTTCGGGGCTGTCATTGTCTGTTTGTCATTCAAGGGCCCAAGCTGAAGAGGGGGT
>gi|175387629|ref|PM_001716884.1| Its the code again
AGACAAGCTTGTCCTGATGTTCCTTGCCCTGGCAGATGTTCAGGACCTTCCTTTGATTCAACCCCTCCAC
CTAAATGGCCCAAGCTTTCGGGGCTGTCATTGTCTGTTTGTCATTCAAGGGCCCAAGCTGAAGAGGGGGT
AAGCTTTCTAAATGGCCCCGGGGCTTTGTCATTCAAGGGTGTGCCCAAGCCATTGTCTTGAAGAGGGGGT
CGGGGCTTTTCAAGGGTCATTGTGCCCAAG
>gi|192669295|ref|XM_001716884.1| Its the code 
AGACAAGCTTGTCCTGATGTTCCTTGCCCTGGCAGATGTTCAGGACCTTCCTTTGATTCAACCCCTCCAC
CTAAATGGCCCAAGCTTTCGGGGCTGTCATTGTCTGTTTGTCATTCAAGGGCCCAAGCTGAAGAGGGGGT
END_File2
#        1        2          3         4         5         6         7
#234567890123456789012345678901234567890123456789012345678901234567890

my $rsFile2out = \ do{ my $dummy };
open my $outFH, q{>}, $rsFile2out
   or die qq{open: > "File2out.txt": $!\n};

my $record = <$in2FH>;
while( <$in2FH> )
{
    if( m{^>gi} )
    {
        processRecord( $record );
        $record = $_;
    }
    else
    {
        $record .= $_;
    }
}
processRecord( $record );

close $in2FH or die qq{close: < "File2.txt": $!\n};
close $outFH or die qq{close: > "File2out.txt": $!\n};

print ${ $rsFile2out };

sub processRecord
{
    my $record = shift;
    my $key = ( split m{\|}, $record )[ 1 ];
    if( exists $positions{ $key } )
    {
        my( $header, @data ) = split m{\n}, $record;
        my $dataStr = join q{}, @data;
        substr
           $dataStr,
           $positions{ $key }->{ position } - 1,
           1,
           $positions{ $key }->{ location };
        print $outFH join qq{\n},
           $header,
           unpack( q{(a70)*}, $dataStr ),
           q{};
    }
    else
    {
        print $outFH $record;
    }
    
    return;
}
[download]

The output.

%positions = (
               '175387629' => {
                                'location' => '*',
                                'position' => '127'
                              },
               '169212695' => {
                                'location' => '*',
                                'position' => '7'
                              },
               '155369268' => {
                                'location' => '*',
                                'position' => '5'
                              }
             );
>gi|155369268|ref|NM_001100917.1| After all
AAAC*ATGTCGATTCTATGATGCGAACGCAGCATTTCAGGGACTGGATGAGGAGCTTACGGTTTTTTACT
ACAGAATCATCAATATCTTGGAAGAAAAAGAATGTTAAGAAATAACAAAACAATAATTATTAAGTACTTT
>gi|169212695|ref|XM_001716884.1| Its the code 
AGACAA*CTTGTCCTGATGTTCCTTGCCCTGGCAGATGTTCAGGACCTTCCTTTGATTCAACCCCTCCAC
CTAAATGGCCCAAGCTTTCGGGGCTGTCATTGTCTGTTTGTCATTCAAGGGCCCAAGCTGAAGAGGGGGT
>gi|175387629|ref|PM_001716884.1| Its the code again
AGACAAGCTTGTCCTGATGTTCCTTGCCCTGGCAGATGTTCAGGACCTTCCTTTGATTCAACCCCTCCAC
CTAAATGGCCCAAGCTTTCGGGGCTGTCATTGTCTGTTTGTCATTCAAGGGCCCAA*CTGAAGAGGGGGT
AAGCTTTCTAAATGGCCCCGGGGCTTTGTCATTCAAGGGTGTGCCCAAGCCATTGTCTTGAAGAGGGGGT
CGGGGCTTTTCAAGGGTCATTGTGCCCAAG
>gi|192669295|ref|XM_001716884.1| Its the code 
AGACAAGCTTGTCCTGATGTTCCTTGCCCTGGCAGATGTTCAGGACCTTCCTTTGATTCAACCCCTCCAC
CTAAATGGCCCAAGCTTTCGGGGCTGTCATTGTCTGTTTGTCATTCAAGGGCCCAAGCTGAAGAGGGGGT
[download]

I hope this is helpful.

Cheers,

JohnGG

Update: Fixed issues with TABs in posted code mucking up indentation.

Have you ever considered using a Database or/and a Template system for this? I'd look at the DBI module and Template::Toolkit for this. After that it's a simple regex based on the hash (which I'm sure at least 3 regex solutions will be posted by the time I'm finished writing this).

use open IO => ':raw';
[download]