added _pre_store_userdata, remove_bby_analysis_id and remove_by_feature_id functions for user uploads

registry and not get_XXXAdaptor() explicitly.

'name_cs_idx').  The patch (patch_49_50_d.sql) is correct.

The script has been updated accordingly.

 Now we loop through all the xrefs to find an xref of the type known/knownxref and not just the display_zref. Also the routine that moves xrefs form one ensembl object to another (e.g. from transcript to gene) has been rewritten as this was not behaving

added 2 new columns in the table: canonical_transcript_id and canonical_annotation and the respective methods in the API to get/set this new information

  my ( $current_start, $current_end );
as these variables were being declared twice