NEWS: TPP

Version 1.0.0:
  
  o First release of the TPP package for analyzing Thermal Proteome Profiling
    experiments.

Version 1.1.0:
  o Changes in devel will be characterized by version number 1.1.x

Version 1.1.1:

  o bugfixes in config table import from .csv files: 
    1. can now handle different types of delimitors (";", "\t", ",")
    2. automatically removes empty rows after import

Version 1.1.2:
  o the recent update of the gridExtra package to Version 2.0.0 required a small fix to 
    ensure sucessful plotting of the melting curves and parameter tables.

Version 1.1.3:
  o fixed problem in QC-plot generation due to the gridExtra package update: 
    adapted height parameter in grid.arrange

Version 1.1.4:
  o Curve parameters "a" and "b" are now reported in the output of analyzeTPPTR and tpptrCurveFit.

Version 1.2.5:
  o CCR-part: 
      fold changes can now be normalized to their lowest temperature during 
      import by the function tppccrNormalizeToReference. 
      This ensures that the transformed values will always be between 0 and 1.
      When normalizing to the lowest concentration, and this value is 0, a small 
      constant (1e-15) is added to prevent division by zero.
  o CCR-import:
      Argument nonZeroCols can be NULL if no additional filtering is needed.
  o CCR output: 
      To make filtering easier, the 'passed_filter' column now shows 
      FALSE even if proteins could not be used for fitting (instead of NAs).
  o CCR output: 
      To make filtering easier, the 'passed_filter' column now shows 
      FALSE even if proteins could not be used for fitting (instead of NAs).
  o CCR output: 
      column with normalization results was renamed from 'normalized' 
      to 'median_normalized' to distinguish from the newly introduced normalization 
      to value at lowest concentration.
  o Excel export: 
      Columns are only color-coded by experiment, when the number of
      experiments is > 1.
  o Excel export: 
      Columns are only color-coded by experiment when the number of
      experiments is > 1.
  o Excel export:
      Relative paths to the plots work now for TR- and CCR part
  o Bug fix for data import: 
      Unique identifiers now treated correctly again.
  o Bug fix for Excel output: 
      Boolean column entries are only transformed to  "yes"/"no" for non-missing values.
  o Bugfix in TR-QC plots: 
      do not attempt to create Tm difference histograms if only one experiment is provided.
  o Bugfix in TR normalization: 
      fixedReference argument works again
  o Bugfix in TR-QC plots: 
      do not attempt to plot minSlope vs. Tm-diffs if no valid Tm-diff values are available.

Version 2.1.0:
  o Major update to the CCR-part: It is now possible to fit and plot multiple
    experiments simultaneously.
  o It is now possible to perform user-specified comparisons of different 
    experiments. They are specified in the 'comparison' column of the config 
    table.
  o TR-part: Hypothesis testing is now separated from result table creation. Therefore, 
    a new function was introduced (tpptrAnalyzeMeltCurves) -> see vignette.
  o CCR-part: Curve fitting and plotting are now conducted by separate functions -> see
    vignette.
  o Bug fixes
  o Introduced color coding of the columns belonging to different experiments in 
    the excel output. This requires openxlsx version >= 2.4.0.
  o The CCR workflow now only returns normalized measurements if normalization was 
    actually performed. Unmodified measurements are always returned and indicated 
    by the suffix 'unmodified'.
  o Data import from tab-delimited files now ignores quotes so that protein annotation 
    fields can contain single ' or " characters.
  o Now enabling arbitrary numbers of plot colors for melting curves or dose response 
    curves.

Version 2.1.1:
Updated package vignette.
  
Version 2.1.2 (Nov-19-2015):
  o Bugfix in config table import:
      All non-alphanumeric characters in the Experiment column are now replaced 
      by'_'.
 o Bugfix in QC scatter plots:
      Can now handle cases where protein IDs do not overlap between experiments,
      or where the temperatures differ between experiments. The affected plots 
      are left empty and contain a message to explain the problem.
      
2.1.3 (Dec-17-2015):
 o Added column 'pEC50_quality_check' to CCR output to filter out pec50 values 
 not within concentration ranges.
 
2.1.4 (Dec-23-2015)
 o Removed example dataset containing fitted data from the Data folder in order 
   to reduce the overall package size (it was only used by some examples in the 
   function Documentation). 
 o Minor changes to documentation in order to ensure that R CMD check and 
   BiocCheck run again without warnings.
   
2.1.5 (Jan-04-2015)
 o Now selectively importing functions from grid and gridExtra package to  
   prevent 'double import' warning due to recent changes in ggplot2.
   
2.1.6 (Jan-18-2016)
 o New CCR output column 'pEC50_quality_check' now reports pec50 values with
   the same sign as the 'pec50' column.
 o Remove importFrom(grid, unit), because unit() is alreay imported by ggplot2 
   (version 2.0.0.), which caused a warning. 
   Added dependency ggplot2 (>= 2.0.0) to ensure that grid::unit is still available.
   
2.1.7 (Jan-22-2016)
 o Bugfixes required after recent ggplot2 update:
    - Included selective import of the grid::gTree function for QC plots.
    - Fixed problem with QC-plot histograms to re-activate plotting of contours in addition to areas.

2.2.1
Update to Bioconductor Release 3.3

2.2.2 (Apr-16-2016)
Removed unit test that causes R CMD check to crash  since the latest update of package 'testthat'.

2.2.3 (Jun-02-2016)
 o Bugfixes in the functions responsible for filtering by user-specified QC columns (e.g. 'qssm')  for normalization set creation. It is now possible to leave the 'otherRequirements' slot in the normalization criteria empty.
 o Fixed typos in function documentation of 'analyzeTPPTR()'
 o Fixed bug in plotColors for the case when no comparisons are specified
 
2.2.4 (Sept-14-2016)
 o Bugfix in 'tpptrCurveFit': restrict file names to 255 characters when creating
   melting curve plots in order to prevent file system crashes.
   
2.2.5 (Oct-04-2016)
 o Bugfix in CCR workflow: avoid division by 0 when transforming foldchanges 
   before dose-response curve fitting.

2.2.6 (Oct-06-2016)
 o Bugfix in CCR workflow: convert filter command back to subet command.