Broaden the definition of complete choroplasts

[greatfireball]

The definition of a complete chloroplast comprises the following requirements:

1. The subgraph need to have between MINNODES (3) and MAXNODES (100)
   

   fastg-parser/import.pl

   Line 200 in be99085

   next if (@{$wcc} < $MINNODES || @{$wcc} > $MAXNODES);

2. Need to be a cyclic subgraph with a total sequence length between MINSEQLEN (25 kbp) and MAXSEQLEN (1 Mbp)
   

   fastg-parser/import.pl

   Line 213 in be99085

   next unless ($c->is_cyclic && $seqlen >= $MINSEQLEN && $seqlen <= $MAXSEQLEN);

3. Subgraph need to have at least one blast hit against the reference database
   

   fastg-parser/import.pl

   Lines 233 to 239 in be99085

   	my $output = qx(tblastx -db $blastdbfile -query $filename -evalue 1e-10 -outfmt 6 -num_alignments 1 -num_threads 4);
   	if (length($output) > 0)
   	{
   	$L->debug("Found hits for cyclic graph: ".$c);
   	push(@cyclic_contigs_with_blast_hits, $c);
   	}

4. Only one subgraph having blast hits is allowed
   

   fastg-parser/import.pl

   Line 246 in be99085

   if (@cyclic_contigs_with_blast_hits == 1)

5. The node with the highest connectivity is assigned as IR
   

   fastg-parser/import.pl

   Line 270 in be99085

   my $inverted_repeat = "$degree[0]{v}";

6. After removing the IR nodes, only two other nodes are allowed
   

   fastg-parser/import.pl

   Line 281 in be99085

   if (keys %nodes == 2)

7. LSC and SSC are simply assigned by sequence length
   

   fastg-parser/import.pl

   Lines 286 to 289 in be99085

   	if (length($seq[$lsc]) < length($seq[$ssc]))
   	{
   	($lsc, $ssc) = ($ssc, $lsc);
   	}

I think we can improve our detection by avoiding some of those requirements, eg. 6.

Any ideas are welcome!