This is really neat++.

As the target is genome work which usually involves large volumes of big strings, using split '[a-z]+', string; to avoid the capture brackets will save a little time.

sub delirium{
    my( $string, $qRef ) = @_;
    $string =~ tr[A-Z][a-z];
    $string =~ s/$_/uc $_/ieg for @$qRef;
    return reduce{ 
        length $a > length $b ? $a : $b 
    } split '[a-z]+',;
}
[download]