=== DRomics : Dose response for omics ===

Version 2.2-1 
=============
NEW FEATURES
- An argument facetby2 was added to bmdplotwithgradient()
and to curvesplot() to be able
to split those plots in rows AND columns using facet_grid().
- Add of the function trendplot() to plot the repartition of dose-response
trend per group of items.
- Add of the function sensitivityplot() to plot various summaries of BMD
values per group of items.
- Add of the function bmdplot() that takes extendedres as a first argument as
trendplot(), bmdplotwithgradient(), sensitivityplot() and curvesplot().
- Removing of the drcfit() argument sigmoid.model that was confusing for some users and not useful for a common use of the package.

Version 2.2-0 
=============
NEW FEATURES
- The second-order Akaike criterion (AICc) recommended to prevent overfitting
with small number of data poinst in each dose-response crives was implemented
and defined as the default option for the argument information.criterion in drcfit().
- The example file of the package (?DRomics) was replaced by a vignette to help
the use of the package and of the Shiny application.
- Improvement of the computation of low BMD values for designs with a high
ratio between the maximal and minimal (non null) tested doses, with the add
of two arguments to bmdcalc, minBMD and ratio2switchinlog.
- Add of two columns in the output of bmdcalc (BMR.zSD and BMR.xfold)
- Add of a function plotfit2pdf() to plot all fits (or residual plots)
in a pdf file, using the raw scale or the log scale of the dose and 
removing of the option saveplot2pdf of drcfit().
- Replacement of the class 'metabolomicdata' by the class 'continuousomicdata'
and add of a function continuousomicdata() that is called by metabolomicdata()
but could be used on other types of continuous omics data such as proteomics data.
- default color changed for bmdplotwithgradient() (green replaced by blue for 
color blind people)
- Removing of three of the four datasets from Zou et al. 2017 to make the 
package lighter
- Add a test on residuals for heteroscedasticity and an output of drcfit: 
residualtests

BUG FIXES
- gestion in RNAseqdata() of all the cases for which vst() may give a stop message.


Version 2.1-0 
=============
NEW FEATURES
- Add of the function bmdplotwithgradient()
- Add of the function ecdfquantileplot()
- Add of an argument named information.criterion in drcfit() to choose 
the use of AIC or BIC for the best fit model selection process.
- Add of the possibility to enter data as an R object of class data.frame
- Add of published datasets (Zhou et al. 2017, Larras etal. 2020) and 
corresponding help pages
- Add of function continuousanchoringdata and modification of itemselect()
to enable the selection of significant responses on continuous anchoring data.
- Add an argument dose_log_transfo in plot.drcfit to enable the use of
log tranformation of the x-axis.
- Add an element in the list of drcfit() output : unfitres giving some 
information on selected items for which the modelling step is not successful
- Add of a function to plot raw data on target items optionally with fitted curves 
for items which have be selected in step 2 and for which step 3 was successful
(new function targetplot()).
- Add of the argument transfo.blind in RNAseqdata() 
- Add of the argument free.y.scales in curvesplot() to enable free y scales in facets
and dose_log_transfo to use an x log scale for plot and calculation of signal.
- Add of examples in DRomics.Rd to help a multi-omics approach
- Add of the argument round.counts to enable rounding of read counts that would come
from Kallisto or Salmon.

BUG FIXES
- make the direct use of varianceStabilizingTransformation() automatic
for small RNAseq data sets (low number of items: < 1000) to fix a bug in RNAseqdata() that occured when using vst() with small datasets.


Version 2.0-1 
=============
NEW FEATURES
- Add of a filter in itemselect(), to exclude from the selection items
with a too high proportion of non detected values (assuming they were
imputed to a common minimum value).
- Add of an argument point.type that enables the change of point type
in, ecdfplotwithCI or its coding by a given factor.
- Add of an argument plot.type in function plot.drcfit() to enable
residual plots.

Version 2.0-0 
=============
NEW FEATURES
- Replacement of the function omicdata() by the function 
microarraydata() and add of two new data importation functions,
RNAseqdata() and metabolomicdata().

Version 1.1-3 
=============
NEW FEATURES
- Replacement of the argument named bytypology of plot.bmdcalc
by an argument named by which can taka three values,
"none", "trend", "model" or "typology".
- Add of a function to plot fitted curves (new function curvesplot()).

Version 1.1-2 
=============
NEW FEATURES
- Add of a function to plot the distribution of a variable as an ecdf plot, with confidence 
intervals on this variable (new function ecdfplotwithCI)

Version 1.1-1 
=============
NEW FEATURES
- Add of bootstrap computation of confidence intervals on benchmark doses (new function bmdboot)
- Add of a function to plot the distribution of a variable as an ecdf plot, with confidence 
intervals on this variable (new function ecdfplotwithCI)


Version 1.0-1
=============
NEW FEATURES
- Add of column yextrem in the results of drcfit (y value at the extremum for biphasic curves)

Version 1.0-0
=============

- Initial release.
