You are doing correctly:
$out[45] = "VUS";
if you want to guarantee that the 46th element of @out has the string "VUS". push will only add to the end of the array, whatever that might be next as element is concerned.

$out[45] = "VUS"; pretty much assigns it and there is not way you have "NULL" after that unless it gets changed or your understanding of what you are looking at is not correct.
Do, you want to test it?

Code:

$out[45] = "VUS";
print "$out[45]\n";

Anything I might say now, it is not a criticism of your posted code, but rather trying to understand why you do it.

Code:

my @colsleft = map "null",(1..$#left);
my @colsright = map "null",(0..$#right);

I think you are trying to create two arrays of certain size with some empty value. But what you are creating is two arrays with each element holding the string "null" which it has no meaning in Perl as empty. In Perl, the equivalent would be undef
Nevertheless, you do not need to worry about that. If you create an array and manually add two elements but not in order, the remaining elements are created with the undef value assigned to it.

Example:
I am going to create an array named @a and populate only the third element $a[2] and the eleventh element $a[10]

Code:

perl -MData::Dumper -e '@a[2] = "Third element of a"; @a[10]="Eleventh element of a"; END{print Dumper \@a}'

Take a look at the representation of that array courtesy of the Data::Dumper module:

Code:

$VAR1 = [
          undef,
          undef,
          'Third element of a',
          undef,
          undef,
          undef,
          undef,
          undef,
          undef,
          undef,
          'Eleventh element of a'
        ];

Let me point to a few issues with your posted code.

Code:

    for (@mutations) {
            ($gene,$transcript,$exon,$coding,$aa) = split/\:/; # this takes col AB and splits it at colons
            grep {$transcript eq $_} keys %nms or next;
    }

I do not know how may times the for loop is being executed because it depends of the size of @mutations set previously by my @mutations=split/,/,$vals[9], however I know that the work it does is in vain, since $gene,$transcript,$exon,$coding,$aa will only keep the last iteration of it. The rest of them are overwritten through the loop.

Code:

grep {$transcript eq $_} keys %nms or next;

Is not producing much since the result is not saved anywhere.

Code:

   $out[2]=$gene;
    $out[3]=$nms{$transcript};
    $out[4]=$transcript;
    $out[15]=$coding;
    $out[17]=$aa;
    $out[45]=$classification;

Remember, many of these will only contain the last iteration from the for loop.

Hopefully, I have given you something to consider.