# NEWS #
## 0.99.0 ##
Port `SingleCellAssay` to use `SummarizedExperiment` under the hood.

## 1.0.0 ##
Release for Bioconductor 3.4

## 1.0.1 ##
Doc updates

## 1.0.2 ##
Fix a bug in which `summary,ZlmFit` was calculating waldTests rather than lrTest

## 1.0.3 ##
Improvements and bug fixes to `thresholdSCRNACountMatrix`

## 1.0.4 ##
Add `convertMASTClassicToSingleCellAssay` to convert objects from MAST versions <= 0.99.0

## 1.0.5 ##
Fixup `stat_ell` to deal with changes introduced in ggplot 2.2.1

## 1.1.6 ##
Base class `SummarizedExperiment0` -> `SummarizedExperiment`

## 1.5.3 ##
Transition to use `SingleCellExperiment` and manage log counts slot.


## 1.11.2 ##

- Deprecated functions in version 1.8.0 are now Defunct.  `filter` (defunct since 1.8.0) has been removed.  Use mast_filter now.
- Refactored the `FromMatrix` constructor so that list-like assays don't make a roundtrip through an array.   This allows sparse-matrix and HDF5-backed assays to be provided to the constructor.
- zlm gains an `exprs_value` argument to allow the assay to be selected that will be tested.

## 1.13.6

- deviance_residuals_hook had dependencies on lm.influence internals that abruptly changed in R 4.0 with no obvious fix.  This function is now defunct.
- Defunct functions `cData`, `fData`, `zlm.SingleCellAssay`, `exprs`, `combine` have been removed.

## 1.17.2

- Fix issue in bootstrapping in which some factor levels aren't sampled in a the bag

## 1.17.6

- `exprs_values` was not actually being used to select the assay in zlm.
- `exprs_values` was not being properly propagated to `ebayes`, `lrTest`, `pbootVcov1` and `bootVcov1`.  Added a slot to ZlmFit to manage this. Thanks to IrSoler for the report.

## 1.21.2

- `REML = FALSE` (previous default) for mixed models can now be overridden with `fitArgsC`
- `lrTest` and `summary` gain `parallel` argument