GitLab

changed selenocystein implementation to use SeqEdits and be more consistant with Transcript implementatin

altered sub version so it doesn't not set the value for no good reason

Translation and TranslationAdaptor start_exon and end_exon to start_Exon and end_Exon

DBEntryAdaptor check to see if gene has a stable id before trying to used it to get info

Added Gene caching to improve slice performance (can still be improved more); Fixes to gene creation from lite database

The main focus of the debugging has been to implement the get_all_Genes method.

For genes with transcripts containing non-sticky exons then things work fine:
 - get gene
 - get all transcripts
 - get all exons
 - transform exons from raw to assembly coordinates
 - re-jiggle the exons to which the translation points
 - re-jiggle the exons to which the transcript points
 - check that everything translates nicely

Sticky exon'ed transcripts are however not playing nicely.  This may be a
feature of the test database - one of the contigs on which the sticky exon
lies may be wrong.  The translations of the current test sticky exon is just
plain wrong.  Retrieving subseqs on the exons are wrong to begin with -
pointing therefore at the contig seq.  Needs debugging / checking in here.
The re-jiggling of translation start/end points for sticky exons needs checking
too.

The other major related changes are in the seq methods.  These now return
dna as txt rather than a primary seq object.  This makes calls such as
seq->seq redundant now.  This change has yet to be checked across all modules
for consistency yet.

The only modelue with any significant changes other than that is Transcript where moltype
has been changed to alphabet due to another change in Bioperl 1.0

tests will not work as I havent' changed the dumps yet.

Changed the way genscans are stored and retrieved

Clone
 - removed Feature_Obj and Gene_Obj use statements

Exon
 - inherits from Bio::EnsEMBL::SeqFeature to use the extra gubbins
 - added temporary_id as for Transcript and Gene

SeqFeature
 - debug

Transcript
 - debug

Translation
 - added temporary_id

ContigI
 - removed string checking - takes *far too long for big contigs

FeatureAdaptor
 - debug on inserts

RawContig
 - removed Feature_Obj and Gene_Obj use statements - changed to AnalysisAdaptor and FeatureAdaptor
 - changed analysis to analysisprocess in several methods
 - removed Feature_Obj - changed to AnalysisAdaptor
 - changed the way genscans are fetched.

Utils
 - added fset2transcript_3frame method

StaticContig
 - changed analysis to analysisprocess
 - gets analysis by AnalysisAdaptor and not Feature_Obj
 - gets genscans differently

Major change, most tests broken. Dont update main trunk, consider using branch
for important stuff.