set dbID and adaptor for stored genes before storing xrefs; pass DBEntryAdaptor->store a Transcript object rather than a dbID

Bug fix - the same factor was being imported multiple times if more than one feature was associated with it.

Now each factor is only stored once and all features associated with it point to a single instance.

macaque data

added the commented out line which enables the non caching on the sql side. Uncomment this out to run timing tests / speed comparisons without the cache invalidating it

Changed dbname arg to dbpattern, to allow for use of regular expressions to specify a number of databases to run on sequentially.

Finds frameshifts (modelled as introns of length 1,2,4 or 5 bp) and adds transcript attributes (code 'Frameshift') for them.

modified same codon calculation to allow more than 1 allele in the same position from different source

optional extra parameter of the sequence so that if it is passed the sequence is not reloaded for dumping

Readded MIM to get the proper descriptions. Also redid all the orders to get more of a gap between sources so that it is easier to add new sources later

Assign projected xrefs to genes, transcripts or translations depending on what type of object other xrefs of this type are assigned to.