CHANGES IN VERSION 1.7.03 (2012-03-22)
o fixed missing documentations
CHANGES IN VERSION 1.7.01 (2012-01-16)
o added remove.duplicates option to pint.match
CHANGES IN VERSION 1.5.35 (2011-10-13)
o Fixed bugs with GenomeModels accessors
CHANGES IN VERSION 1.5.34 (2011-08-15)
o changed the default match.probes to TRUE in screen.cgh.mrna; now
also checks with match.probes = FALSE that the probes are matched
CHANGES IN VERSION 1.5.33 (2011-08-02)
o now artificially adding arm information if it is missing, to avoid
halts
CHANGES IN VERSION 1.5.30 (2011-07-15)
o screen.cgh.mrna option segmented changed into match.probes
CHANGES IN VERSION 1.5.29 (2011-05-09)
BUG FIXES
o Fixed a bug in plot.ChromosomeModels
CHANGES IN VERSION 1.5.28 (2011-05-09)
BUG FIXES
o fixed a bug which prevented screeningn while missing values were present
CHANGES IN VERSION 1.5.06 (2011-04-16)
o added interpretation tools in report.R
CHANGES IN VERSION 1.4.02 (2011-04-16)
o still correcting critical bug in fixed.window and associated
functions: now accepting data with no chromosomal arm information
CHANGES IN VERSION 1.4.01 (2011-04-14)
o corrected critical bug in fixed.window and associated functions
CHANGES IN VERSION 1.3.85 (2011-04-11)
o Added a method to get the latent variable Z for particular model
CHANGES IN VERSION 1.3.84 (2011-04-10)
o vignette updated
CHANGES IN VERSION 1.3.83 (2011-04-10)
BUG FIXES
o Multiple bugs with GenomeModels methods fixed
CHANGES IN VERSION 1.3.82 (2011-04-10)
BUG FIXES
o plot.ChromosomeArmModels now prints by default the arm in the
plot title if only models from one arm is plotted
CHANGES IN VERSION 1.3.80 (2011-03-27)
SIGNIFICANT USER-VISIBLE CHANGES
o Dependency modeling has been splitted to dmt package.
o ChromosomeArmModels class is removed from the package.
CHANGES IN VERSION 1.3.71 (2011-03-10)
o The sign of the effects in z.effects and W.effects now corresponds with the data
CHANGES IN VERSION 1.3.70 (2011-03-04)
BUG FIXES
o fixed bug which caused an error when using findModel to return a model which is
in the end of a chromosome
CHANGES IN VERSION 1.3.69 (2011-03-02)
o added as.data.frame method
BUG FIXES
o plot method for GenomeModels now works with showTop argument
CHANGES IN VERSION 1.3.68 (2011-03-02)
o corrected bibtex bug in doc
CHANGES IN VERSION 1.3.67 (2011-02-01)
o if row or column names for the data are missing, they are named by
indices and warning is given (pint.data function)
CHANGES IN VERSION 1.3.66 (2011-01-29)
BUG FIXES
o pint.data: dimnames now provided for the output data and info fields
CHANGES IN VERSION 1.3.63 (2011-01-13)
BUG FIXES
o as.data.frame fixed
CHANGES IN VERSION 1.3.63 (2010-01-11)
o Added many new development features that help to handle segmented
data (options segmented and nonmatched, and related scripts). Will
be documented properly in later versions.
CHANGES IN VERSION 1.3.35 (2010-12-30)
BUG FIXES
o in pint.match added rownames check between Y$info and
Y$data, previously only number of rows was checked for Y data. For
X data things were ok: both row number and row names were checked
o in pint.data call within pint.match, the X data set was
erroneously used for both X and Y calls. Now calling Y for Y and X
for X. Also polishing replace.inf part. Now also naming the info field
rows correctly.
NEW FEATURES
o checking info fields in pint.data for synonymous terms: replacing
"chromosome" with "chr" and "stop" with "end". Requires less
tweaking from the end user.
o if chromosome names X or Y used, replace these with 23 and
24. This makes code compatible with any organism.
MISC
o changed the internal function name 'impute' to 'imputation'
o moved some functionality from pint.match to pint.data
CHANGES IN VERSION 1.3.33 (2010-12-09)
BUG FIXES
o corrected bug in impute.R function which produced NAs
o added replace.inf in pint.data; now this preprocessing function
(and consequently pint.match) replaces infinite values with large
values by default
CHANGES IN VERSION 1.3.32 (2010-12-09)
NEW FEATURES
o imputation option added (default TRUE) to pint.match and improved
in pint.data
CHANGES IN VERSION 1.3.3 (2010-10-21)
NEW FEATURES
o GenomeModels and ChromosomeModels classes have getModelNumbers
method
o topGenes and topModels methods have default values
o screen.cgh.mrna checks whether data has been segmented and
warns the user
CHANGES IN VERSION 1.3.2 (2010-10-20)
NEW FEATURES
o Default value for windowSize argument in screening function
changed
CHANGES IN VERSION 1.3.1 (2010-10-19)
NEW FEATURES
o Segmented gene copy number data can be screened with pSimCCA
method, although it is not recommended.
o Output type of screening functions can changed to a data frame.
o All dependency model outputs from screening functions can now
converted to a data frame with as.data.frame method.
o ChromosomeArmModels can be accessed with [['p']] and [['q']]
from ChromosomeModels.
BUG FIXES
o z.effects gives a proper error message when the window size is
too large.
o pCCA method gives a proper error message when the data matrix is not
invertible.
CHANGES IN VERSION 1.1.17 (2010-10-15)
HELP & Rd FILE CHANGES
o sample and genes are displayed in correct order in help pages
for screening functions
CHANGES IN VERSION 1.1.16 (2010-08-04)
BUG FIXES
o screening functions can handle chromosomes with fewer samples
than window size
o isEmpty method is now exported
o topModels method for ChromosomeModels class gives no longer
an error when either arm has no dependency models
CHANGES IN VERSION 1.1.15 (2010-06-16)
SIGNIFICANT USER-VISIBLE CHANGES
o updates to pint.match function. Now confirms that 'data' and
'info' fields in the input data matrices are converted to have
matching elements (corresponding to geneids, for example). Also an
option to include only selected chromosomes was added. Works now
with both X/Y or 23/24 annotation for the chromosomes; characters
and numerics are both ok. Changes affect also the screen.mrna.cgh
function which calls pint.match (and other functions that call it)
CHANGES IN VERSION 1.1.1 (2010-06-04)
NEW FEATURES
o matches the samples between cgh and mrna data (screen.cgh.mrna
function) automatically based on sample names
o matches genes between cgh and mrna data (screen.cgh.mrna function)
based on chromosomal locations. User-defined distance threshold
can be given.
BUG FIXES
o added robustness towards missing values
CHANGES IN VERSION 0.99.5 (2010-03-28)
SIGNIFICANT USER-VISIBLE CHANGES
o screen.chromosome function name changed to screen.cgh.mrna
o methods findHighestGenes and findHighestModels changed to topGenes and topModels
o z.projection and W.projection changed to z.effects and W.effects
o method name 'TPriorpSimCCA' removed and added to 'psimCCA'
NEW FEATURES
o screening between mirna and cgh added
o ability to show gene names with highest dependency scores added to plots
o model can be searched by gene name from output from screen functions
o model can be calculated around gene by its name
o order.results methods added which return genes and their dependency scores
o size of variable name texts can be changed in plot.DependencyModel
o aliases added to methods (pfa, pcca, pcca.isotropic and ppca)
o function to format data for packages dependency model calculation added (pint.data)
o function to pair data sets added (pint.match)
HELP & Rd FILE CHANGES
o clarified the documentation
DOCUMENTATION
o model description added to Vignette
CHANGES IN VERSION 0.99.3 (2010-02-23)
SIGNIFICANT USER-VISIBLE CHANGES
o calculate function name changed to screen.chromosome
NEW FEATURES
o probabilistic factorial analysis model added
o argument M in fit.dependency.model changed to H and arbitary matrix
can be given instead of just identity matrix for pSimCCA method with T prior
o plot.DependencyModel function added
o z.projection and W.projection functions added.
HELP & Rd FILE CHANGES
o changes to examples in help pages
DOCUMENTATION
o Vignette updatet
o Collate field added to DESCRIPTION
BUG FIXES
o all methods of fit.dependency.model are fixed to give full W and phi
matrices with correct dimension names
CHANGES IN VERSION 0.99.4 (2010-03-08)
SIGNIFICANT USER-VISIBLE CHANGES
o default method of screen.chromosome() changed to pSimCCA
NEW FEATURES
o psimCCA method now accepts data matrices of a different size
o used method in screen.chromosome() can be determined by parameters only
o annotation types can be given to plot.DependencyModel() function
o samples can be ordered in plot.DependencyModel() function according to their
contribution to dependency model
o dimension of latent variable z can be given with psimCCA method
BUG FIXES
o bug fixed that prevented screening the whole genome
o show methods now print the used method when only parameters given to screen.chromosome()
o incorrect gene name in DependencyModel objects corrected