my $r = {};  # hash to hold the parsed record
while (<IN>) {
  if (/^(\d+).../) { # found beginning of new record
    if ($r->{id}) { process($r); }
    $r = {};         # begin new record
    $r->{id} = $1;   # populate parsed info from this line
  } elsif (/KEGG.../) {
    $r->{kegg} = ...;
  } elsif ...
  }
}
if ($r->{id}) { process($r) };

sub process {
  my $r = shift;
  ...
}