RnBeads 1.7.1 ============= * Several improvements in age prediction module * Several minor bug fixes RnBeads 1.5.1 ============= * Reference-based cell type composition estimation now uses the 50,000 most variable sites by default. * Minor bug fixes RnBeads 1.3.8 ============= * Added age prediction from 450K and bisulfite data to inference module RnBeads 1.3.7 ============= * Subsampling for reporting and plotting methylation values distributions in filtering report to reduce memory footprint. RnBeads 1.3.6 ============= * Support for Illumina EPIC array and related bugfixes * Added a section on whitelist and blacklist in the Preprocessing module RnBeads 1.3.5 ============= * Fixes with ggplot in 450K array QC plotting * Added an option 'differential.report.sites' to decrease runtime by skipping the differential site methylation section in the report RnBeads 1.3.4 ============= * Added support for Illumina EPIC array RnBeads 1.3.3 ============= * Added 'smooth.profile' parameter to rnb.plot.locus.profile function RnBeads 1.3.2 ============= * Performance and memory improvements during loading, qc and preprocessing of large bisulfite-derived RnBSets * Added optional arguments to meth() and covg() allowing for partial retrieval of methylation and coverage information from RnBSets. Particularly makes sense for large, disk-based datasets. * Added option "disk.dump.bigff.finalizer" for setting the finalizer for temporary BigFfMat objects RnBeads 1.3.1 ============= * Added "remove.regions" function for RnBSet class * summarize.regions now summarizes for each sample individually in order to reduce memory consumption for large WGBS datasets * import.skip.object.check option for keeping the memory profile low while loading huge datasets * NA sites (filtering.missing.value.quantile) are now removed even if they were previously masked (filtering.low.coverage.masking) * Added "nsites" method for quickly extracting the number of sites for large RnBSet objects without having to retrieve the full methylation matrix * Added "hasCovg" method for quickly determining whether coverage information is present in large RnBSet objects without having to retrieve the full coverage matrix RnBeads 1.1.9 ============= * Filtering report fix when no normalization is conducted * Bugfixes for combine(RnBSet, RnBSet) and BigFf matrices RnBeads 1.1.8 ============= * Corrected coverage statistics in sample summary table: Sites with NA methylation values are no longer considered in the coverage statistics (makes a difference if some coverage threshold is applied) * Improved method for gender prediction. Predicted genders are also included in the exported annotation table RnBeads 1.1.7 ============= * Improvements to mergeSamples function for RnBiseqSets * Some more memory clean-up RnBeads 1.1.6 ============= * Differential methylation based on region level only is now supported * Minor updates to the differential methylation report generation * Performance improvements and minor bugfixes for using disk.dump.bigff * Performance improvements (more memory clean-up) RnBeads 1.1.5 ============= * New option: disk.dump.bigff for using a wrapper class of the ff package to avoid their INT_MAX issue RnBeads 1.1.4 ============= * Fixes in parallel environment setup from bioconcuctor RnBeads 1.1.3 ============= * Some fixes in data loading * Fixes in parallel environment setup RnBeads 1.1.2 ============= * New annotation package format * Support for filtering out cross-reactive probes in Infinium 450k dataset * Improved logging on a Mac RnBeads 1.0.0 ============= * Bioconductor release RnBeads 0.99.23 =============== * Incorporated some bioconductor check related improvements RnBeads 0.99.22 =============== * Incorporated bioconductor requirements for parallel processing. "parallel.disable" is now "parallel.teardown" RnBeads 0.99.20 =============== * RnBeads no longer suggests IlluminaHumanMethylation450k.db. RnBeads 0.99.19 =============== * Fixed a bug in the computation of surrogate variables (SVA) that did not take into account other adjustment variables in the model (compared to the null model) * Updated locus profiles to work with the latest version of Gviz * Differential methylation reports now contain information on group sizes * Included plots of number of sites vs coverage percentiles for each sample in the dataset (rnb.plot.num.sites.covg) * Bisulfite-seq QC reports now contain an additional section with summary statistics and plots when the coverage histogram plots are enabled RnBeads 0.99.18 =============== * Implementation of class constructors (the S4 constructors are not available anymore) * Analyses can now be started from an RnBSet which is stored on the hard drive (using save.rnb.set) via the rnb.set.dir data type * New IDAT loading routine, based on the illuminaio package * Multiple fixes in wrappers for wateRmelon normalization methods * Fixed the RefFreeEWAS wrapper * Updated the Houseman, 2012 reference-based method implementation * RnBSets now know the version with which they were created * Various other bugfixes and improvements to reports and documentation RnBeads 0.99.17 =============== * Consistency issues with Bioconductor 3.0 fixed * Fixed a bug relating to no coverage masking being conducted for sequencing data * Various other bugfixes RnBeads 0.99.16 =============== * Enhanced XML-based analysis though the tag preanalysis.script * Optimized loading of the sequencing data sets * Enhanced region profiles plots. Added region site distribution plots * Locus profiles can now be generated in the exploratory report for regions specified in custom bed files and for genes listed as gene symbols * Improved cell type heterogeneity inference (no cell types excluded anymore) * Added more support for paired differential methylation with limma * Some functions were renamed: merge.samples->mergeSamples, add.pheno->addPheno, add.region.subsegments->addRegionSubsegments, and all functions operating on Report * Added helper classes for submitting RnBeads to a scientific compute cluster. Current implementation includes Sun Grid Engine * The default pipeline now uses ff functionality and saves intermediate objects into the report directory * Minor bugfixes RnBeads 0.99.15 =============== * Restructuring of the pipeline modules: ** the loading module has been renamed to import ** the prefiltering,normalization and postfiltering modules have been summarized in a new module: preprocessing ** the batch and profiles modules have been summarized in a new module for exploratory analysis ** the export module has been renamed to tracks and tables ** corresponding option names have been changed ** see the overview figure on the website or the vignette for a quick overview on the new module structure * Multiple RnBSets can now be concatenated with the add() function * Multiple samples in an RnBSet can be merged using the merge.samples() function * Gender prediction can be performed on Infinium 450k datasets * Minor updates on covariate adjustment * Updates on tissue heterogeneity estimation * Calling differentially methylated sites with RefFreeEWAS now supports paired design * Multiple minor bugfixes and performance improvements * Vignette and documentation updates RnBeads 0.99.13 =============== * Added new module on annotation inference * Added correction for cell type heterogeneity * Added SVA functionality * Minor bugfixes * Updates to the vignette and other documentation RnBeads 0.99.12 =============== * Support Bismark coverage file loading * Enhanced documentation and logging of loading steps * Minor bugfixes * Updates to the vignette and other documentation RnBeads 0.99.11 =============== * Accommodate data packages for individual genomes * Performance: disk space usage when using disk dumping, options for subsetting sites when computing distance matrices for clustering, PCA and MDS * Region subsegmentation * The default method for differential methylation p-values is limma RnBeads 0.99.10 =============== * Performance improvements * Option to keep big matrices on the hard drive rather than main memory * Restructuring of the filtering modules. Parts of the filtering steps are executed before normalization others afterwards * Improvements in normalization: more methods supported * Cosmetic changes to some of the plots RnBeads 0.99.9 ============== * Bugfixes in paired analysis * Usability of bisulfite sequencing RnBeads 0.99.8 ============== * New normalization methods integrated * Improved arguments to rnb.run.analysis * Improved parsing of the sample annotation table * Paired analysis (testing stage) * Webservice installed * Multiple bugfixes RnBeads 0.99.7 ============== * Support for background subtraction and BMIQ normalization of Infinium 450k data * Support for differential methylation analysis on all pairs of sample groups RnBeads 0.99.6 ============== * Locus Profiles * Support for parallel computing RnBeads 0.99.0 ============== * Many additional features including bisulfite sequencing mode, the mouse genome, data export, ... RnBeads 0.5.0 ============= * Initial release of RnBeads containing six modules - loading and normalization, quality control, filtering, batch effects, methylation profiles and differential methylation.