CHANGES IN VERSION 1.2.2 -------------------------- o bug fix #50: correct peaks-matrix, to handle xcms bug (sneumann/xcms#220) for older xcms-versions (thx to @lauzikaite) o test order of parameters to optimize CHANGES IN VERSION 1.2.1 -------------------------- o bug fix regarding conflict of BPPARAM and nSlaves arguments (thx to @lauzikaite) CHANGES IN VERSION 1.2.0 -------------------------- o Bioconductor release 3.5 CHANGES IN VERSION 1.1.2 -------------------------- o vignette updated o plot margins omitted CHANGES IN VERSION 1.1.1 -------------------------- o use package BiocParallel (via argument `BPPARAM`) instead of `nSlaves` to controll `xcms`-parallelization o depends on xcms >= 1.50.0 o formatting `writeRScript` to output more beautifully CHANGES IN VERSION 1.1.0 -------------------------- o merge Bioconductor 1.0.0 release code with Github code. CHANGES IN VERSION 1.0.0 -------------------------- o Biocondcutor release 3.4 ========================== BIOCONDUCTOR release ========================== After version 1.7.5 IPO was put on Biocondcutor. After that the versioning schema changes to meet the Bioconductor version numbering and starts with 1.0.0. CHANGES IN VERSION 1.7.5 -------------------------- o added usage of clustering method FORK on unix-systems (thanks to Pablo Moreno) o fixed bug in parallelization to prevent conflicts with package 'snow' CHANGES IN VERSION 1.7.4 -------------------------- o preceded parallel-functions with 'parallel::' to use right package o fixed bug in function writeRScript using 'loess' retention time cor. o decreased runtime for R CMD check IPO CHANGES IN VERSION 1.7.3 -------------------------- o added runnable examples o decreased size of pictures in vignettes/rsmDirectory o decreased runtime for unit-tests o replaces expand.grid with expand.grid.subset (in utils.R) CHANGES IN VERSION 1.7.2 -------------------------- o bugfix: try to prevent error in calcPPS possibly caused by NAs o replaced cat() and print() calls with message() CHANGES IN VERSION 1.7.1 -------------------------- o checking correlation of peak-shape with sinus curve (-pi/2 to pi*1.5), normal distribution or checkBorderIntensity o findIsotopes.IPO renamed parameter checkBorderIntensity to checkPeakShape o performance improvement calcPPS for checkPeakShape=FALSE o calculating xcmsSet-object and respective PPS for each DoE. (PPS is not estimated from rsm anymore) o additionally forwarding nSlaves for xcmsSet-function (also see getDefaultXcmsSetStartingParams()) CHANGES IN VERSION 1.7.0 -------------------------- o added support for XCMS-method retcor.loess o updated help files o changed return value of getRGTVValues o adapted unit tests o parameter scanrange for XCMS-methods findPeaks can be set but not optimized CHANGES IN VERSION 1.6.2 -------------------------- o Updated the function getNormalizedResponse() to prevent NAs CHANGES IN VERSION 1.6.1 -------------------------- o Added installation script and installation description in vignette CHANGES IN VERSION 1.6 -------------------------- o Added support of CAMERA isotope identification (findIsotopes.CAMERA) o selectivity of findIsotopes.IPO may be increased if checkBorderIntensity is set to TRUE: 'maxo' of each peak of an isotopologue must be three times higher than the intensities at 'rtmin' and 'rtmax' o simplified return value of calcPPS() to vector with meaningful names o changed getDefaultXcmsSetStartingParams() for min_peakdwidth = c(12, 28) and for ppm to c(17, 32) CHANGES IN VERSION 1.5.7.1 -------------------------- o using predict() to identify best levels and expand.grid to generate testdata for model CHANGES IN VERSION 1.5.7 -------------------------- o supporting single parameter optimization. Only basic version with redundant levels in consecutive DoEs o removed integer-rounding in maximum focusing for all findPeaks parameters except prefilter(I) and steps o Update documentation to match code o Remove use of getwd() preventing absolute subdir paths CHANGES IN VERSION 1.5.6 -------------------------- o updated vignette o generally using Central-Composite design instead of Box-Behnken design o fixed bug when defining subdir=NULL in functions optimizeXcmsSet and optimizeRetGroup o modified unit tests to handle versions > 1.5.6 o added function writeRScript to NAMESPACE export o updated man for optimizeXcmsSet and optimizeRetGroup CHANGES IN VERSION 1.5.5 -------------------------- o increased recall of reliable peaks in calcPPS. Changed exponent for PPS calculation from 1.5 to 2 CHANGES IN VERSION 1.5.4.8 -------------------------- o remove file lookup in optimizeXcmsSet, and leave that to xcmsSet() CHANGES IN VERSION 1.5.4.7 -------------------------- o explicitely use serial evaluation if nSlaves=1, to help debugging CHANGES IN VERSION 1.5.4.6 -------------------------- o remove dependency on Rmpi CHANGES IN VERSION 1.5.4.5 -------------------------- o added examples from msdata o fix Depends, imports and library() and require() calls CHANGES IN VERSION 1.5.4.4 -------------------------- USER VISIBLE CHANGES o packaged script o changed method name attachparams to attachList o changed method name calculateRGTV to getRGTVValues o changed method name getDefaultStartingXcmsParams to getDefaultXcmsSetStartingParams o changed method name typeCastFactor to typeCastParams o changed method name writeRSkript to writeRScript o changed the parameter name n_slaves to nSlaves o resultIncreased: if last optimization score was 0, no isotopes have been found hence the dataset is not optimizable with IPO. o added man files for attachList, calcPPS, combineParams, createModel, decode, decodeAll, encode, getBbdParameter, getCcdParameter, getDefaultRetCorCenterSample, getDefaultRetGroupStartingParams, getDefaultXcmsSetStartingParams, getNormalizedResponse, getRGTVValues, IPO-package, optimizeRetGroup, optimizeXcmsSet, startSlaves, toMatrix, typeCastParams o removed xcmsSetsettingsAsString.R o getResponses: now able to handle NULL value for slices parameter o calcPPS: peaks with NA values are removed before isotopes identification o if subdir is NULL, no rsm's are saved o optimizeXcmsSet: lowere minimum value for min_peakwith from 5 to 3 #IPO_V1.5.4.3: * LIP calculation in calcPPS fixed #IPO_V1.5.4.2: * added initial parameter check # * renamed all factor-variables to params # * in optimizeXcmsSet: - also look for mzML-files # - check if files were found # * bug in optimization for matchedFilter fixed; sigma and mzdiff have to be # definded later (combineFactors()) when sigma and step as well as steps are already known #IPO_V1.5.4.1: changes in calcPPS: # rt_window <- rt * 0.005 # rt_lower <- part_peaks[,"rt"] - rt_window # rt_upper <- part_peaks[,"rt"] + rt_window #IPO_V1.5.4: if bad_group == 0; bad_group = 1 && good_group += 1 #IPO_V1.5.3: no parameter for isotope detection. # c13_peak[,"mz"] has to be within (mzmin + isotope_mass) and (mzmax + isotope_mass) # c13_peak[,"rt"] has to be within (rtmin + isotope_mass) and (rtmax + isotope_mass) #IPO_V1.5.: in RCSandGSIncreased: also used good_groups ^ 2 #IPO_V1.4.: vectorized isotope identification; # no intensity window, between intensity of max carbon and 1 #IPO_V1.3.: good_groups ^ 2 to increase recall