Version 1.3.3
________________________________________________________________________________
- Use homegrown Rcpp based graph class and algorithm for clique extraction

Version 1.1.12
________________________________________________________________________________
- Bug fix: pangenome subsetting is now faster and actually correct

Version 1.1.11
________________________________________________________________________________
- Improvement: Various speed improvements
- Fix namespace clash between Matrix and S4Vectors

Version 1.1.10
________________________________________________________________________________
- Feature: New class pgSlim for handling pangenomes with no ref to sequence data
- Bug fix: safeAAread/safeDNAread would return wrong sequences when number of 
fasta files exceeded 2000

Version 1.1.9
________________________________________________________________________________
- Feature: Threshold for core group classificaton can now be set (defaults to 1)
- Improvement: Only investigate the neighbors to groups that have changed during
iteration in neighborhoodMerge.

Version 1.1.8
________________________________________________________________________________
- Bug fix: Vignette error due to problems with reutils

Version 1.1.7
________________________________________________________________________________
- Feature: cdhitGrouping creates initial grouping based on cdhit algorithm
- Feature: neighborhoodSplit now refines the splitting as a final step by 
merging highly similar groups sharing gene group up- or downstream
- Feature: gpcGrouping can now precluster using CD-Hit
- Feature: Key algorithms now reports progress and timing information
- Feature: Custom linearKernel function that takes an upper similarity threshold
to speed up comparisons.
- Feature: Updated vignette, focusing on recommended workflow
- Improvement: More performant pcGraph, neighborhoodSplit, pgMatrix methods
- Improvement: pgMatrix now returns a sparseMatrix for lower memory footprint
- Improvement: pangenome matrix no longer stored in pgInMem
- Bug fix: Remove zerolength genes upon pangenome creation.
- Bug fix: Batch accessing fasta files to avoid "too many open connections" 
error

Version 1.1.6
________________________________________________________________________________
- Skips some long running examples and tests to cut down on check time

Version 1.1.5
________________________________________________________________________________
- Bug fix: .fillDefaults now correctly extracts the frame in advance

Version 1.1.4
________________________________________________________________________________
- Fixing imports to import BiocGenerics, S4Vectors and IRanges in full

Version 1.1.3
________________________________________________________________________________
- Minor optimization of code
- getRep now names genes by group name
- transformSim now works on sparseMatrix rather than matrix objects - avoids 
coercing huge sparse matrices down to matrix format

Version 1.1.1
________________________________________________________________________________
- Fix bug that would affect pangenomes above 100 genomes when GPC was used with
automatic tree creation.

Version 1.1.0
________________________________________________________________________________
- Added kmerSplit method

Version 1.0.3
________________________________________________________________________________
- Update to work with ggplot2 v2.0.0
- All imports from ggplot2, gtable and grid are now explicit

Version 1.0.2
________________________________________________________________________________
- Fixed bug in transformSim that would cause it to break on large similarity
matrices
- Fixed bug in orgTree only present with pangenomes over 100 genomes

Version 0.99.0
________________________________________________________________________________
- Submission to Bioconductor
