Changes in Version 1.27.2

   o segment function gets default.DNAcopy.bdry internaly instead of assuming
     DNAcopy is loaded. Allows for loading package by NAMESPACE only.
   o xmaploc now works for samplebychrom and chrombysample also.


Changes in Version 1.27.1

   o Don't further segment if all values in the current segment are the same.
   o Use is.finite in functions instead of is.na(.) & !(abs(.)==Inf)


1.25.1 became 1.26.0 in Bioconductor 2.8 & =1.27.0 now
------------------------------------------------------

Changes in Version 1.25.1

   o House keeping: changed the URL for cytoband and removed Columbia email


1.23.8 became 1.24.0 in Bioconductor 2.7 & =1.25.0 now
------------------------------------------------------

Changes in Version 1.23.8

   o Added getbdry to exported functions in NAMESPACE

Changes in Version 1.23.7

   o Versions 1.23.4 to 1.23.6 were changes made to use file based data. 
     This needed changing CNA and DNAcopy objects to S4 classes which was
     too much of a paradigm shift for users.  So changes are undone here.
   o Change default smooth.region to 10 in smooth.CNA
   o Added function plotSample for whole genome plot of a single sample.
   o Fixed the is.integer to is.numeric in subset functions.
   o CNA accomodates data.frames (as.matrix not needed) and presorted data.

Changes in Version 1.23.3 (same fix applied in v1.22.1)

   o Fixed weird bug in smoothCNA.f triggered when there are at most
     2*smooth.region (default smooth.region=2) markers in a chromosome.

Changes in Version 1.23.2

   o row numbers of the segments are automatically returned as the object
     segRows in the output.  The print method shows it if showSegRows=TRUE.

Changes in Version 1.23.1

   o Added zoomIntoRegion function.
   o Added an option to return the row numbers of the segments.


1.21.6 became 1.22.0 in Bioconductor 2.6 & =1.23.0 now
------------------------------------------------------

Changes in Version 1.21.6

   o Fixed indexing bug (i <= k or i > n-k, not and) in smoothCNA.f

Changes in Version 1.21.5

   o ADDED packageStartupMessage THAT CNA DATA FORMAT WILL CHANGE
   o changed filename of NEWS to CHANGES; file format didn't work with news
   o changed function .First.lib to .onLoad
   o in the function CNA, changed code for ordering by chrom and maploc 
     because of "na.last=NA" slowed it down terribly
   o removed the unnecessary rownames command in CNA

Changes in Version 1.21.4

   o moved CHANGES file to inst/NEWS
   o documentation of CNA and segment say that missing values will be removed

Changes in Version 1.21.3

   o forgot to deallocate memory in the new code

Changes in Version 1.21.2

   o changed the code for htmaxp & hwtmaxp for the hybrid

Changes in Version 1.21.1

   o changed our affiliations in the vignette


1.19.6 became 1.20.0 in Bioconductor 2.5 & =1.21.0 now
------------------------------------------------------

Changes in Version 1.19.6

   o added function glFrequency to calculate gains and losses

Changes in Version 1.19.5

   o Changed smooth.CNA from all R to Fortran backbone

Changes in Version 1.19.4

   o Fixed infinite loop when the logratio is constant across a segment

Changes in Version 1.19.3

   o Redundancy check by Henrik Bengtsson
   o Bioconductor fixes (license version, BiocViews etc.)

Changes in Version 1.19.2

   o Fixed a rarely triggered bug in the new code


Changes in Version 1.19.1

   o Faster algorithm to compute the CBS maximum t-statistic


1.17.5 became 1.18.0 in Bioconductor 2.4 & =1.19.0 now
------------------------------------------------------

Changes in Version 1.17.5

   o For weighted CBS return weighted segment means (not ordinary mean).

Changes in Version 1.17.4

   o Bug fixes to wtermp code and how it's called in wfindcpt.

Changes in Version 1.17.3

   o Weighted segmentation code added.  Weights per probe can be used.

Changes in Version 1.17.1

   o NAMESPACE file added and man pages fixed to reflect it.


1.15.4 became 1.16.0 in Bioconductor 2.3 & =1.17.0 now
------------------------------------------------------

Changes in Version 1.15.4

   o make sure there are at least 10 obsns when t-stat threshold is used

Changes in Version 1.15.3

   o the minimum width of segment is user selectable (argument min.width). 
     Changed from the fixed value of 2 based on user demand.  Max value 5.

Changes in Version 1.15.2

   o changed to a faster algorithm for calculating the max t-statistic

   o changed to a faster code for undoing edge effects (redundant perm)

   o added basic functions exon.segment and exon.changepoint (workhorse)
     These functions are used for finding a translocation using binary
     segmentation.  The data for this should be properly normalized.

   o fortran code re-arranged into different files

Changes in Version 1.15.1

   o updated the call to get(getOption("device")) to use dev.new


1.13.3 became 1.14.0 in Bioconductor 2.2 & =1.15.0 now
------------------------------------------------------

Changes in Version 1.13.3

   o Added segments.summary to give the median, sd & mad for each segment.

Changes in Version 1.13.2

   o Modified the p-value function to replace p > 1 with 1.

Changes in Version 1.13.1

   o Modified the p-value function.  Added pseudo confidence intervals 
     for the change-points.

   o Windowing option has been removed (R & fortran code modified).


1.11.2 became 1.12.0 in Bioconductor 2.1 & =1.13.0 now
------------------------------------------------------

Changes in Version 1.11.2

   o Added a p-value function for the change-points.  This is based on 
     binary segmentation and not necessarily correct but gives an idea
     of the relative importance of the change-points.

Changes in Version 1.11.1

   o Changed Venkat's affiliation from MSKCC to Columbia University


1.9.3 became 1.10.0 in Bioconductor 2.0 & =1.11.0 now
-----------------------------------------------------

Changes in Version 1.9.3

   o Added warning that windowing will be deprecated in the next version.

Changes in Version 1.9.2

   o Added code to not bother with p-values and split the segment when the
     maximal T is large (p-value will be tiny if T > 7).

   o Added code to not split when the maximal T is small (fixes a numerical
     problem where a constant large segment can have a significant split).

Changes in Version 1.9.1

   o The mod function in fortran inflicts serious time penalty --
     code rewritten to fix it


1.5.3 became 1.6(8).0 in Bioconductor 1.8(9) & =1.9.0 now
---------------------------------------------------------

Changes in Version 1.5.3

   o Fixed the subset functions to enable sample re-ordering

Changes in Version 1.5.2

   o Bug fix in plot.DNAcopy (triggered only when #chrom=1)

   o Added information about stopping rule in the Vignette

   o Cytoband data from the goldenPath repository added for future plots

Changes in Version 1.5.1

   o Add a sequential stopping rule to declare change early 

   o Bug fix - make the object from subset.DNAcopy of class DNAcopy

   ###############################################################
   ###                                                         ###
   ##      E. S. Venkatraman is the new package maintainer      ##
   ###                                                         ###
   ###############################################################



1.2.5 became 1.4.0 in Bioconductor 1.7 & =1.5.0 in the development branch
-------------------------------------------------------------------------

Changes in Version 1.2.5

   o Updated the vignette to include references to additional features.

Changes in Version 1.2.4

   o Seg fault occurs if window.size is set when hybrid method is used.
     Fixed by setting window.size to be NULL in the function segment.

   o Added options to plot command to draw line at y=0 and control its 
     color and lwd of all lines.

   o If ylim is missing it's calculated from all the data instead of just
     the first sample.

Changes in Version 1.2.3

   o Modifications to the plot function to make par("ask") behave better.

Changes in Version 1.2.2

   o Modified the plot function.  X-axis is either the index or (cumulative) 
     map location. Incorporate col, pch, ylim etc. to provide better control
     over plots.

Changes in Version 1.2.1

   o Added the hybrid method for computing p-values to determine splits.  This
     speeds up the computations considerably and so has been made the default.

