CHANGES IN VERSION 1.37.6
--------------------------
NEW FEATURE:
     o Introducing write.mzQuantML(xcmsSet) to export the peak list and grouped matrix to the PSI format mzQuantML (see http://www.psidev.info/mzquantml)
USER VISIBLE CHANGES:
     o Add Brigham Young University to LICENSE file for copyright purposes. 
     o Add copyright information display when running findPeaks.massifquant()
     within xcmsRaw.R
     o Clean and update documentation for findPeaks.massifquant-methods.Rd  
BUG FIXES:
     o Remove unused parameters in findKalmanROIs() within xcmsRaw.R
CHANGES IN VERSION 1.37.5
--------------------------
BUG FIXES
    o fixed bug in retcor.obiwarp where the scanrange of the first sample would be checked instead of the center sample

CHANGES IN VERSION 1.37.4
--------------------------
BUG FIXES
    o Skip t-test in diffreport() if one class has less than 2 samples.

CHANGES IN VERSION 1.37.3
--------------------------
BUG FIXES
    o fixed bug in patternVsRowScore (group.nearest) that was introduced by the 
      modifications in rev 65169 and caused features to be aligned that were far outside the given m/z and retention time windows.


CHANGES IN VERSION 1.37.1
--------------------------
BUG FIXES
    o fixed fillPeaks, which 1) dropped non-standard columns 
    and 2) failed if nothing to do, based on patches by Tony Larson.

CHANGES IN VERSION 1.37.1
--------------------------

NEW FEATURES:
     o Introducing msn2xcmsRaw, to allow findPeaks() on MS2 and MSn data


CHANGES IN VERSION 1.35.7
--------------------------
BUG FIXES
    o fixed indexing bug in group.nearest, which under certain circumstances caused all peaks in the first sample to be ignored (reported by Tony Larson)

CHANGES IN VERSION 1.35.6
--------------------------
BUG FIXES
    o Obiwarp retention time alignment error-ed if scanrange was used as a parameter setting during xcmsSet/peak detection
        The method now tries to automatically find the set scanrange and uses this range for alignment.

CHANGES IN VERSION 1.35.4
--------------------------

NEW FEATURES:
     o Introducing parallel fillPeaks

USER VISIBLE CHANGES
    o Replace snow requirement with minimum R version 2.14.0


CHANGES IN VERSION 1.35.3
--------------------------
BUG FIXES
  o if group.density was used with very low minfrac settings (< 0.5) it did not return all feature groups, 
    but only those that include features from at least 50% of samples in a group. This limitation was removed.


CHANGES IN VERSION 1.35.2
--------------------------

UPDATED FEATURES:
     o Behind the scenes xcms now uses the xcmsSource class to read raw data.
       This allows e.g. to write a class that pulls raw data 
       from e.g. a database
BUG FIXES
     o massifquant: simplified logic structure of Tracker::claimDataIdx
     resolved failure on new test case. 
     o massifquant: reporting features data structure compatible with 
     multiple sample comparison within XCMS.

CHANGES IN VERSION 1.35.1
--------------------------

UPDATED FEATURES:
     o The mzData export is now much faster and uses less memory

CHANGES IN VERSION 1.33.16
--------------------------

USER VISIBLE CHANGES

    o diffreport and plotEIC have a new parameter mzdec, with is the number of decimal places
      of the m/z values in the EIC plot title

CHANGES IN VERSION 1.33.16
--------------------------

UPDATED FEATURES:
Lock mass gap filler now works with netCDF lock mass function file to find the exact times of the scans and works with
the newer Waters MS instruments.

CHANGES IN VERSION 1.33.15
--------------------------

BUG FIXES

    o scanrage is now honoured in xcmsSet, also when in parallel mode

CHANGES IN VERSION 1.33.14
--------------------------

BUG FIXES

    o scanrage is now honoured in xcmsRaw, and consequently
      also in xcmsSet(matchedFilter), where previously
      it was ignored.

CHANGES IN VERSION 1.33.13
--------------------------

BUG FIXES

    o write.cdf() has been fixed to write files AMDIS can read

CHANGES IN VERSION 1.33.12
--------------------------

BUG FIXES

    o write.mzData adds Polarity to the file if available

CHANGES IN VERSION 1.33.11
--------------------------

USER VISIBLE CHANGES

    o centWave uses a new method to estimate local noise which improves detection of closely spaced peaks

NEW FEATURES

    o placeholder

BUG FIXES

    o group.mzClust was failing when result had one peak


For more details and all changes before May 2012 please see the (now discontinued) CHANGELOG in the source package (inst/ folder).

CHANGED BEHAVIOUR since Version 1.32:

Other Changes since Version 1.32:

* improved mzData writing, now includes MSn spectra and less verbose.
* improved netCDF writing, but not yet good enough for AMDIS


CHANGED BEHAVIOUR since Version 1.14:

* centWave may report a smaller set of peaks, due to a small bug
  in the ROI algorithm some features with mass deviation > ppm were retained.

Other Changes since Version 1.14:

* New method for grouping: an algorithm inspired by mzMine
  group(method="nearest") has been implemented. It is slower
  than group(method="density"). It can individually group
  close-eluting peaks of very similar mass

* New method for retention time correction:
  The retcor(method="obiwarp") algorithm operates on the raw data,
  and thus allows to correct runs without well-behaving
  peak groups, or without peak picking at all.

* fillPeaks(method="MSW") is now also available
  for direct infusion spectra. The findPeaks(method="MSW")
  now returns several intensities, and correctly reports
  mzmin and mzmax for peaks.

* centWave now uses dynamic memory allocation, needs much less memory,
  and these BUF related errors should be a thing of the past.

* centWave gains an optional argument "noise",
  which is useful for data that was centroided without any intensity threshold,
  centroids with intensity < "noise" are omitted from ROI detection

* the fillPeaks() methods now remember which was
  an observed, and which was a "filled" peak.

* For direct infusion spectra diffreport() now shows
  the raw peak shapes, and also indicated "real" and "filled" peaks.

* xcmsRaw can now filter for positive/negative spectra,
  if the file includes both polarities. xcmsSet() can pass
  the polarity to contain positive/negative peaks only.
