to fetch ungapped_features from DNA-PEP transcript_supporting_features
(due tothe presence of Insertions that are non modulo-3 in length, which is
common for introns).

Added -delete_unmapped option to only delete stuff from the unmapped_object table if the option is specified. This makes it easier to do "patch" xref runs without clobbering the unmapped data from a previous full run.

Added temporary(?) OTTT source for Havana-Ensembl transcript mappings

try to fall back to gene analysis when storing transcript that doesn't have an analysis object attached

Some sources (e.g. TIGR Gene Index) reuse accessions between species. I have therefore altered the "insert_or_select" and "get_xref_id_by_accession_and_source" functions so that they accept a species arg. The UNIQUE key on the xref table has also been updated with the addition of the species field

Add join with seq_region to allow output of chromosome name when -locations option is used. Seems to slow things down, so could be removed again if -locations is removed in future.

Now throws if a dependent_xref is found, but the source (EntrezGene|MIM) are not in the xref DB. Without this change we end up with sql execute errors later on

During initial source selection, now selects the species in addition to the species_id. If the source_url is mysql:* the source and species are fed to the parser as additional args to the run method. Species and source specific queries can thus be generated

change to throw if dbpattern isn't found