Avoid duplicating data when anonymising it for export to external analysis tool

The file IO routines strictly rely on metadata in SequenceI when generating IDs and sequence names in the files they emit, which means these fields must be renamed prior to export in order to ensure ID strings are unique and do not contain weird characters.

Rather than have to create a copy of data to be output, the SeqHash mechanism could be embedded within the file IO system to allow unique ID generation and recovery in-situ rather than before output and after import as is currently done.