Modifed get_all_DBLinks to sort xrefs by priority (desc), external db name (asc), display_id (asc) before returning them.

dbprimary_acc_linkable (column 5): default true (1) EXCEPT HUGO
display_label_linkable (column 6): default false(0) EXCEPT HUGO

priority: all set to 1 initially.

allow MArkersymbol sources to be used to get the gene descriptions. If all fails use Marker symbol as gene description regardless of filters.   Still work in progress

Only store descriptions from certain sources (listed in gene_description_sources()) to prevent e.g. IPI being used for description if there's nothing else found.

Fix to method of finding transcript display xrefs; previously xrefs from same source would not be reliably sorted on query/target identity.

allows the removal of only certain types of seq region attributes from
a slice)

Added RefSeq_dna_predicted & RefSeq_peptide_predicted to priority list (below known RefSeqs and SPTREMBL).

Slight change to description filtering logic; use original unfiltered description if anything survived the filtering.