only study core, est, cdna and coreexpression databases, better checking of analysis.description file

changed config for elephant

Get the sources to be used for display_xrefs and gene decriptions from the species file if it exists

use priority level as well as the list. Things like MGI and HGNC can come form many places so make sure you get the best one

canonical_transcript() method warn about not being able to fetch the
canonical transcript when there's no adptor attached to the gene.

transcript.  Saves fetching a transcript for every gene.

internal ID of the transcript instead of the transcript itself.  The
canonical_transcript() method will then lazy load the transcript.

This speeds up the gene adaptor code.

The old semantics is maintained.

if script is ran then set checksum to 1. In subsequent parsings test checksum and only run the script if it is not 1

  Make bind_columns() list easier to read.

  Remove mentionings of $external_name as it was always undef.
  Instead set external_name attribute of new gene to undef
  ($gene->external_name() will use display_id from display_xref
  unless external_name is set explicitly).

if there is a canonical transcript attached to the gene.  If there
isn't, store a 0 (zero).  The canonical_transcript_id field does not
allow for NULLs.  This patch makes the unit tests happy.

canonical annotation is stored along with the rest of the gene data.