1.19.5
------
Fix issue with quality assessment vignette not building correctly due to internal changes in NUSE() function


1.19.4
------
Fix additional NAMESPACE issue

1.19.3
------
Added a NAMESPACE



1.19.2
------
change all relevant VECTOR_ELT/SET_VECTOR_ELT to STRING_ELT/SET_STRING_ELT. A number of functions are affected by this change (fitPLM() being a primary example)




1.19.1
-------
fitPLM() with the default model now returns values (coef, se) for the last probe in probeset


1.17.1
------
Fix typo in get.psi.code (Geman-McClure issue)
(see http://article.gmane.org/gmane.science.biology.informatics.conductor/19075)



1.15.9
------
rma background correction code from preprocessCore is now used
rma_background2.c and rma_background2.h are now removed


1.15.8
------
remove weightedkerneldensity.c which was not used anywhere.


1.15.6
------
Replaced usage of soon to be defunct class exprSet with ExpressionSet.


1.15.5
------
Add narrays to PLMset-class, to ensure that show(Pset) works right. Fix summary(Pset)


1.15.4
------
Removed fitPLM.old, PLM.designmatrix, PLM.designmatrix2 which were all legacy defuct functions.


1.15.3
------
Fixed very minor typo in vignette


1.15.2
------
Remove some legacy C code from do_PLMrlm.c. This code was not part of the workflow for any user constructed fitPLM call. Removing it should help improve maintainability.


1.15.1
------
Add a CITATION



1.13.16
-------
Resync with changes in preprocessCore




1.13.13
-------
commit the deletion of a file that should have been removed from
the repository

1.13.12
-------
restructure medianlogPM and logmedianPM code



1.13.10
-------

Continue clean up with code that had previously been moved to preprocessCore




1.13.9
------
Remove R_rlm_interfaces.c from affyPLM (it has been moved to preprocessCore 


1.13.8
------
Fix so vignettes build



1.13.7
------
move affydata from Depends to Suggest:



1.13.6
------
change MAplot vignette to refer to ExpressionSet rather than exprSet objects


1.13.4
------
core rlm code is now in preprocessCore. Also moved AvgLog, LogAvg and lm code there.



1.13.3
------

The core medianpolish algorithm is now from preprocessCore


1.13.2
------
fix a compiler warning



1.13.1
------
remove the qnorm.c code. This has been moved to preprocessCore, along with normalize.quantiles.determine.target(), normalize.quantiles.use.target() and a few related things.





1.11.13
-------
PLMset no longer contains an eSet object and has no direct dependencies on ExpressionSets etc
all the exprSet functions/methods have been replaced with equivalent ExpressionSet methods




1.11.12
-------
Additional fixes to get it to pass R CMD check (which it now does). MAplot for ExpressionSet.
PLMSet will no longer "contain" an eset object from next version (1.11.13)
All exprSet methods/functions will be removed in next version (1.11.13)


1.11.11
-------

Some fixes to make it pass R CMD check. changes how const.coefs are stored.


1.11.9-1.11.10
--------------
These versions where done by people at FHCRC. Basically modernizing the package. Try to make some things that were exprSet dependent instead be eset/ExpressionSet based.


1.11.8
------
Small optimization for RMA background C code.



1.11.7
------
Some small memory/speed optimizations in some of the C code


1.11.6
------
Add SE output to raw rlm function (for direct calling).



1.11.5
------
Fixed an unbalanced UNPROTECT



1.11.4
------

instances of "\." in a pattern were changed to "\\."


1.11.3
------
fix minor verbosity problem in pre-process



1.11.2
------

Introduce preprocess() for general pre-processing (ie normalize and background correcting)



1.11.1
------

Make functions less verbose (this can be controlled with verbosity.level parameter). This applies to fitPLM, threestep, threestepPLM, rmaPLM.
change a few apply(x,1,median) to rowMedians






1.9.17
------
Fix R_moduleCdynload on windows so it should now build


1.9.16
------
Fix a makevars problem.


1.9.15
------
Change moduleCdynload to R_moduleCdynload


1.9.14
------
The model PM ~ -1 + samples + treatment:probes was being misidentified in C code and the default model fitted instead.

1.9.13
------
a small fix to boxplot() when applied to a PLMset containing a non-default model

1.9.12
------
A .Call() at the C level for accessing the optimized rlm for the rma type model.


1.9.11
------
Fix small qnorm find/apply target bug


1.9.10
------
Added hist() for exprSet


1.9.9
-----
Added an additional vignette explaining various ways to use the MAplot function.
Removed the subset argument from MAplot (for exprSet, PLMset)
Added groups argument to MAplot


1.9.8
-----
add ability to calculate scaling factors for scaling normalizations on the log2 scale. This modification of the scaling
normalization method was recommended by   Lu, Chao (2004) Improving the scaling normalization for high-density oligonucleotide GeneChip expression microarrays. BMC Bioinformatics 5:103


1.9.7
-----
Modify MAplot for exprSet, PLMset objects for changes in ma.plot from affy.


1.9.6
-----
Added normalize.AffyBatch.quantiles.chromosome



1.9.3-1.9.5
-----------

Small fix to fitPLM for soybean arrays.
A number of quantile normalization variants.




1.9.2   May 1, 2006
-----

ReadRMAExpress() can deal with gzipped files.



1.9.1   Apr 27-28, 2006
-----

Better checking of validity of normalization method and normalization parameteres supplied to fitPLM, threestep etc.
quantiles.robust is now a supported normalization option (to match with normalize.quantiles.robust in the affy function)







very OLD CHANGELOG is below:


1.4.1
-----

Fix for bug in fitPLM when model is PM ~ -1 + samples + probes (ie default model) when there are probesets with only single probeset
Changed default setting of varcov from "all" to "none" in output.param in fitPLM ) this reduces memory over head.


1.4.2
-----

Boxplot on a exprSet by now uses range=0 by default


1.4.3
-----

Fix a potential gc() problem in threestep
Fix threestepPLM() so it works on Windows.
