Sorted out possible SQL precedence problem (idea from James).

  Old SQL (fragment):

    FROM (transcript t)
      LEFT JOIN transcript_stable_id tsi
        ON tsi.transcript_id = t.transcript_id
      LEFT JOIN xref x
        ON x.xref_id = t.display_xref_id
      LEFT JOIN external_db exdb
        ON exdb.external_db_id = x.external_db_id

  New SQL (same fragment):

    FROM
      ( ( ( (transcript t)
            LEFT JOIN transcript_stable_id tsi
              ON tsi.transcript_id = t.transcript_id
          )
          LEFT JOIN xref x
            ON x.xref_id = t.display_xref_id
        )
        LEFT JOIN external_db exdb
          ON exdb.external_db_id = x.external_db_id
      )

    Arg [7]    : boolean $no_fuzz
                 If true, do not use "fuzzy matching" (see below).

  This was done to allow the user to by-pass the fuzzy matching code.
  This will allow for some speed increases and to remove any ambiguity
  around the results from this method.

  Also inserted clarifying comment about stickleback DB ('groups' vs.
  'chromosomes') from Anne.

sequence returnes by seq call returns the number of bases specified by the length of the StrainSlice

the coding region does not start/end in this exon, it also will not
end/start here.

host added for pre-sites

options standardised

Small change to build_transcript_and_gene_display_xrefs subroutine so that gene get their own display_xref instead of being assigned one from a transcript.

Made pattern matching more strict when choosing type (gene, transcript, translation) as 'pseudogene' transcripts had previously been classified as genes.

added optional parameter of code to fetch_all_by_ and remove_from_ so only attibutes with these codes are used

added method delete_cache to remove ALL the assembly caching. Needed for when the assembly is changed

Only the ID, Name and FlyBase_Annotation_IDs are given synonyms (from Alias=).
Removed some unused subroutines.

Also specify host etc to use load_registry_from_db() in project_display_xrefs.pl so don't have to specify everything in conf file.

Now use load_registry_from_db() as well as spcifying settings in config file. Should mean that only Compara has t be defined in conf file.

clarified documentation of fetch_all() regarding combining toplevel and coord system version arguments

added support for retrieving features and slices to be projected from a different database; documentation

data

Added

human -> chimp, macaque, guinea_pig
drosophila -> aedes