RELEASE HISTORY OF coseq PACKAGE ======================================== coseq VERSION 1.15.1 - 1.15.6 -- Improved documentation in vignette, especially FAQs. (Thanks to Florian Rocher, David Roux, and Jessica Keffer for questions leading to these clarifications.) -- Fixed non-reproducibility observed when seed argument is provided for Poisson mixture model and K-means with parallel computing. (Thanks to Océane Cassan for the bug report.) coseq VERSION 1.13.1 - 1.13.2 -- Adjust plotting function to remove ggplot2 warning. (Thanks to Praveen Kumar for the pull request.) -- Add option to remove red lines from boxplots with add_lines argument. (Thanks to Marie-Laure Martin-Magniette for the suggestion.) -- Fix behavior when seed is to be set for reproducible results for the Normal mixture model via the new seed argument. coseq VERSION 1.7.1 - 1.7.3 -- Add ImmunoOncology to biocViews in DESCRIPTION -- Add informative error message when a subset of estimated Gaussian mixture models have singular covariance matrices. -- Add generic for plot function. -- Specify UTF-8 encoding in DESCRIPTION. -- Minor updates to vignette and documentation. coseq VERSION 1.5.2 -- Fix use of n_row and n_col in plotting function, and fix bug in coseq plotting when data feature an unbalanced number of replicates per condition. (Thanks to Yuguo Xiao for the bug report.) -- Remove dependency on gridExtra package (previously the use of n_row and n_col In plotting used marrangeGrob, but this is now directly dealt with using ggplot2). -- Update vignette to illustrate how to customize coseq plots via ggplot2. coseq VERSION 1.5.1 -- Clarify in documentation that normalization can be suppressed by setting normFactors = "none" in coseq. (Thanks to David Rangel for many helpful discussions.) -- K-means can now be run for less than 10 models without provoking an error; however, a message appears to inform user that model selection via capushe is not possible (and the largest model is automatically selected). -- Updated citations. -- The coseq plot function now has the argument collapse_rows, which can take the values c("none", "sum", "average"). When collapse_rows = "sum", the plot behaves the same as when average_over_conds = TRUE , and profiles for replicates within a condition are summed. When collapse_rows = "average", values for replicates are averaged; this behavior may be preferable for highly unbalanced designs. (Thanks to Gosia Komor for the idea.) coseq VERSION 1.3.2 -- Added some clarification in the documentation of coseq, especially to clarify that genes should be in rows and samples in columns. -- Added some checks/warnings for “normal” vs “Normal”, applying transformations to profiles -- Updated maintainer email coseq VERSION 1.1.3 -- Corrected error causing failing test for compareARI function when a matrix of non-discrete and/or non-positive values was provided. coseq VERSION 0.99.8 - 0.99.12 -- Add Fietz et al. (2012) data to package -- Add in K-means algorithm functionality and conditional probability calculation. -- Add flexibility in graph options (output ggplot object from function, xlab, ylab, use_sample_names, optional facet labeling) -- Updated vignette and citation info. coseq VERSION 0.99.5 -- Add unit tests, remove static vignette. coseq VERSION 0.99.1 -- Fairly significant changes to address Bioconductor reviewer comments. In particular, the package now makes use of S4 classes and methods: a new S4 class, coseqResults, has been created to extend the RangedSummarizedExperiment S4 class from the SummarizedExperiment package. coseq VERSION 0.1.x - 0.99.0 -- Initial development version of coseq.