# CHANGES IN VERSION 2.11.1

* Updated consensus assembly in consICA(). Components of every run are now
  aligned to the best run with a one-to-one (bijective) matching, and the
  consensus is accumulated from zero. Previously the average was seeded with
  the first run in its native order (its own alignment was discarded), and the
  per-component matching could assign one source to several components. This
  removes a stability drop at ntry=2 and corrects the `stab` metric; results
  for large ntry (e.g. 20) are unchanged.
* saveReport() no longer errors for single-run results (ntry=1, no stability).


# CHANGES IN VERSION 2.0.0

* allowed input format: SummarizedExperiment, matrix, Seurat
* new: ANOVA component ~ factors and plot weigths
* new: calculate similarity of GO lists from different runs
* parallelization for getGO()


# CHANGES IN VERSION 1.3.1

* Speed up consICA() function


# CHANGES IN VERSION 1.1.2

* fix bug with filtering and upd documentation


# CHANGES IN VERSION 1.1.1

* Include GO in cica object and implement auto setOrientation()
* Add setOrientation() function (for positive significant GOs)
* Renamed parameter: 'cica' in all functions
* Add checkers 'compliant to `consICA()` result'


# CHANGES IN VERSION 0.99.1

* fix Notes 
* add unit tests
* vectorization


# CHANGES IN VERSION 0.99.0

* Submit to Bioconductor.