We have two sets of changes here. The first supports the use of alternative stores of sequence data but still providing the sequence caching framework meaning repeated requests for a similar region will not incur a cost of going back to the original data store.

The second change is writing in native Perl support for samtool's FAIDX code. This includes indexing the FASTA files and providing access to the uncompressed files. Our new indexing code is available as a SequenceAdaptor implementation.