I tried running start_site_curation.pl against a single BSML file containing multiple contigs and ran into a few bugs that caused information to leak across contigs.
1. The array for recording genes on a contig was shared across all calls to the sequence handler.
2. The XMLPath for finding gene nodes was from the document root rather than the current element.
3. The fasta parser didn't recognize to stop at the beginning of the next contig.
I've attached a patch containing the changes I made. I'm not familiar enough with the script to be comfortable committing the changes unreviewed.
Thanks,
Chris
start_site_curation.pl diff
Assigning to Kevin, as this is part of his pipeline and I also wouldn't be able to properly gauge the affects of a change.