Add get_Gene() method to CDS, ExtonTranscript and UTR objects to match Transcript method, for consistency in accessing.

Added source_dir and target_dir parameters allowing copy of databases outside the MySQL default internal directory on Sanger machines /mysql/data_3306/databases/. When target_dir is specified, the script will automatically create a symlink from target_dir to /mysql/data_3306/databases/. This change will allow us to copy databases from and to the new ens-staging3 MySQL server

This reverts commit cbd14e9f.

Compara test db for Slice methods

Changing all ens-staging1/2 to ens-staging3, as we'll be using this new host for pipelines until the EBI shift over

It should be seq_region_start and seq_region_end for UTRs, since start/end now are relative to the feature, as with other feature types

Because of a fix of a very old bug in the mapper, some Exon features if they're made via a Slice have start() and end() relative to the feature. So in genomic2cdna calls we were using $self->start() and $self->end(), this means we were sometimes using incorrect coordinates. Change to seq_reion_start and seq_region_end

bugfix: get_all_DBAdaptors() should return an empty list if the species name (alias) does not exist

Tweak to gene description filters

Support for multi letter patches

- Fix how UTRs are created, they should be created with start/end rel…

- Fix how UTRs are created, they should be created with start/end relative to the feature, not in genomic coordinates
- Update tests to check UTRs are created correctly when transcripts are fetched from a Slice
- UTR test cases for forward and reverse strand transcripts

Changes to gtf serializer to deal with offset in codon and utr reginos