v1.3.13 - Moved gtools from Suggests to Imports - Removed multiple instance of 'f' function in ExpressionSetList[ method v1.5.3 - Corrected bug in zeroNu function (needed to pass one.delta argument to empiricalStart function) - fixed bug in reading acceptance and potential for XdeMCMC objects v1.5.4 - added burnin argument to calculatePosteriorAvg function v1.5.5 6/4/2009 - For parameters indexed by genes and platforms, the chains were read in incorrectly ## - The order of the vector for parameter nu_gp indexed by gene g and study p is ## nu_11, nu_12, ..., nu_1P, nu_21, ..., nu_2P, nu_G1, ..., nu_GP ## ## Arrays: the 'left subscript moves the fastest' ##Therefore ## mcmc <- array(mcmc, dim=c(# iterations, # genes, # studies)) is equiv to ## For iteration 1: ## mcmc[1, 1, 1] <- nu_11 ## mcmc[1, 1, 2] <- nu_12 ## mcmc[1, 1, P] <- nu_1P ## mcmc[1, 2, 1] <- nu_21 ## mcmc[1, 2, P] <- nu_2P v1.5.6 6/4/2009 - Fixed bugs in functions calculatePosteriorAvg and calculateBayesianEffectSize v1.5.7 6/4/2009 - The chains were actually read in correctly prior to v.1.5.5. Changes the comments below in methods-XdeMcmc. - Note that we're returning aperm(mcmc) for parameters indexed by genes and platforms, so the changes in "bug" fix 1.5.6 still work ##--------------------------------------------------------------------------- ##Parameters indexed by gene and study ## ## If ITER iterations were executed, there are ITER lines written to the file. ## Each line is a big vector in the following order: ## gene1 study 1, gene 1 study2, ... , gene1 study P, gene 2 study 1, ... ## Or, ##iter1: nu_11, nu_12, ..., nu_1P, nu_21, ..., nu_2P, nu_G1, ..., nu_GP ##iter2: nu_11, nu_12, ..., nu_1P, nu_21, ..., nu_2P, nu_G1, ..., nu_GP ## ## for the nu parameter. ## ## For reading in arrays, the 'left subscript moves the fastest' ## ## scan() reads in the data by rows. ## Therefore, the fastest moving index is study, followed by genes and then the mcmc iteration. ## ## mcmc <- array(mcmc, dim=c(# studies, # genes, # iterations)) is equiv to ## For iteration 1: ## mcmc[1, 1, 1] <- nu_11 ## mcmc[2, 1, 1] <- nu_12 ## mcmc[P, 1, 1] <- nu_1P ## mcmc[1, 2, 1] <- nu_21 ## mcmc[2, 2, 1] <- nu_22 ## ... ## mcmc[P, 2, 1] <- nu_2P ## ... ##--------------------------------------------------------------------------- if(name %in% c("DDelta", "nu", "phi", "sigma2", "delta", "probDelta")){ I <- iterations(x) mcmc <- array(mcmc, dim=length(studyNames(x), nrow(x), iterations(x)), dimnames=list(studyNames(x), featureNames(x), as.character(1:I))) ##Now transform this to an array with dimensions: ## iteration x features x study mcmc <- aperm(mcmc) return(mcmc) } ##-------------------------------------------------- v 1.5.8 Fixed bug in converting the mcmc object to an array v 1.5.9 committed 6/22/2009 o updated the elements of expressionSetList -- elements now have scanDates() slots o fixed documentation errors v 1.5.10 committed 6/25/2009 o fix warning: added seed to XdeParameter-class man page v 1.5.11 committed 7/23/2009 o updated the elements of expressionSetList -- elements now have protocolData slots v 1.5.12 committed 8/10/2009 o try to fix compile errors on gewurz: option authoryear and natbib not working. Trying: \usepackage[authoryear,round,numbers]{natbib} o added seed to man/XdeParameter-class slots o updated vignettes / pdf o added NEWS file v 1.5.13 committed 9/29/2009 o removed authoryear option from natbib in Sweave fileg v 1.5.14 committed 10/02/2009 o removed debugging code from xde.R v 1.5.15 committed 10/02/2009 o resurrected old copy of xde.R v 1.7.1 committed 12/04/2009 o added integrativeCorrelationFilter - returns a TRUE/FALSE matrix for each gene x study-pair o import MergeMaid v 1.9.1 committed 05/18/2010 o changed empiricalStart function so that it is possible to use empirical values for a single study