Merge branch 'master' into ci

* master: (45 commits) Bump bioc devel version Bump bioc release version bump x.y.z versions to odd y after creation of RELEASE_3_7 branch bump x.y.z versions to even y prior to creation of RELEASE_3_7 branch Fix logo in README Set seed in sparseDCEstimate tests and bump ver Bump version to 1.3.5 Bump version number Add SparseDC tests Run checks Add SparseDC to listSims Add sparseDCEstimate Add sparseDCSimulate Add SparseDCParams Rename params metadata to Params for consistency Reset path.from if nGroups changes in SplatParams Colourise listSims Colourise Params print output Improve printing of data.frame parameters Improve mean fitting warning in splatEstMean ...

Merge branch 'master' into ci
d0229362 · Luke Zappia · 86901797 · cd54e3e1 · d0229362 · d0229362
Commit d0229362 authored 6 years ago by Luke Zappia
--- a/DESCRIPTION
+++ b/DESCRIPTION
 Package: splatter
 Type: Package
 Title: Simple Simulation of Single-cell RNA Sequencing Data
-Version: 1.3.1
+Version: 1.5.0
-Date: 2018-01-03
+Date: 2018-05-02
 Author: Luke Zappia
 Authors@R:
    c(person("Luke", "Zappia", role = c("aut", "cre"),
-      email = "luke.zappia@mcri.edu.au"),
+      email = "luke.zappia@mcri.edu.au",
+      comment = c(ORCID = "0000-0001-7744-8565")),
      person("Belinda", "Phipson", role = c("aut"),
-      email = "belinda.phipson@mcri.edu.au"),
+      email = "belinda.phipson@mcri.edu.au",
+      comment = c(ORCID = "0000-0002-1711-7454")),
      person("Alicia", "Oshlack", role = c("aut"),
-      email = "alicia.oshlack@mcri.edu.au"))
+      email = "alicia.oshlack@mcri.edu.au",
+      comment = c(ORCID = "0000-0001-9788-5690")))
 Maintainer: Luke Zappia <luke.zappia@mcri.edu.au>
 Description: Splatter is a package for the simulation of single-cell RNA
    sequencing count data. It provides a simple interface for creating complex
@@ -21,7 +24,6 @@ License: GPL-3 + file LICENSE
 LazyData: TRUE
 Depends:
    R (>= 3.4),
-    scater,
    SingleCellExperiment
 Imports:
    akima,
@@ -35,9 +37,11 @@ Imports:
    matrixStats,
    methods,
    scales,
+    scater (>= 1.7.4),
    stats,
    SummarizedExperiment,
-    utils
+    utils,
+    crayon
 Suggests:
    BiocStyle,
    covr,
@@ -55,7 +59,8 @@ Suggests:
    mfa,
    phenopath,
    BASiCS,
-    zinbwave
+    zinbwave,
+    SparseDC
 biocViews: SingleCell, RNASeq, Transcriptomics, GeneExpression, Sequencing,
    Software
 URL: https://github.com/Oshlack/splatter

--- a/NAMESPACE
+++ b/NAMESPACE
 # Generated by roxygen2: do not edit by hand
-S3method(BASiCSEstimate,SCESet)
+S3method(BASiCSEstimate,SingleCellExperiment)
 S3method(BASiCSEstimate,matrix)
 S3method(lun2Estimate,SingleCellExperiment)
 S3method(lun2Estimate,matrix)
@@ -15,6 +15,8 @@ S3method(scDDEstimate,default)
 S3method(scDDEstimate,matrix)
 S3method(simpleEstimate,SingleCellExperiment)
 S3method(simpleEstimate,matrix)
+S3method(sparseDCEstimate,SingleCellExperiment)
+S3method(sparseDCEstimate,matrix)
 S3method(splatEstimate,SingleCellExperiment)
 S3method(splatEstimate,matrix)
 S3method(zinbEstimate,SingleCellExperiment)
@@ -43,6 +45,7 @@ export(newMFAParams)
 export(newPhenoParams)
 export(newSCDDParams)
 export(newSimpleParams)
+export(newSparseDCParams)
 export(newSplatParams)
 export(newZINBParams)
 export(phenoEstimate)
@@ -53,6 +56,8 @@ export(setParam)
 export(setParams)
 export(simpleEstimate)
 export(simpleSimulate)
+export(sparseDCEstimate)
+export(sparseDCSimulate)
 export(splatEstimate)
 export(splatSimulate)
 export(splatSimulateGroups)
@@ -68,6 +73,7 @@ exportClasses(MFAParams)
 exportClasses(PhenoParams)
 exportClasses(SCDDParams)
 exportClasses(SimpleParams)
+exportClasses(SparseDCParams)
 exportClasses(SplatParams)
 exportClasses(ZINBParams)
 importFrom(BiocParallel,SerialParam)
@@ -127,4 +133,5 @@ importFrom(stats,rnbinom)
 importFrom(stats,rnorm)
 importFrom(stats,rpois)
 importFrom(stats,runif)
+importFrom(stats,shapiro.test)
 importFrom(utils,head)
--- a/NEWS.md
+++ b/NEWS.md
+## Version 1.5.0 (2018-01-02)
+* Bioconductor 3.7 devel
+# Version 1.4.0 (2018-01-02)
+* Bioconductor 3.7 release
+## Version 1.3.6 (2018-04-30)
+* Set seed in sparseDCEstimate tests
+## Version 1.3.5 (2018-04-25)
+* Replace dropout.present with dropout.type in SplatParams
+    * Allows users to set dropout by experiment, batch, group or cell
+## Version 1.3.4 (2018-04-19)
+* Add option to use a normal distribution for library sizes in Splat simulations
+* Add SparseDC simulation
+* Rename params in metadata slot of simulation to Params for consistency
+* Add checks to SplatParams to reset path.from if nGroups changes
+* Improve printing of parameters stored as data.frames
+* Colourise Params print output and listSims output
+* Improve warning message when fitting means in splatEstimate fails
+* Simplify problematic documentation links
+### Version 1.3.3.9010 (2018-04-12)
+* Add option to use a normal distribution for library sizes in Splat simulations
+### Version 1.3.3.9000 (2018-04-12)
+* Replace dropout.present with dropout.type in SplatParams
+    * Allows more control over dropout.mid and dropout.shape
+## Version 1.3.3 (2018-03-27)
+* Fix parameter passing bug in scDDEstimate
+## Version 1.3.2 (2018-01-31)
+* Fix QC names that have been changed in scater
+* Move scater to Imports and add scater version
+* Remove lingering references to SCESets
+* Add more tests
 ## Version 1.3.1 (2018-01-03)
 * Fix error in vignette caused by changes in scater
 ## Version 1.3.0 (2017-10-13)
-* Bioconductor 3.7 devel 
+* Bioconductor 3.7 devel
+## Version 1.2.2 (2018-03-27)
+* Fix parameter passing bug in scDDEstimate
 ## Version 1.2.1 (2017-11-23)

--- a/R/AllClasses.R
+++ b/R/AllClasses.R
@@ -95,9 +95,13 @@ setClass("SimpleParams",
 #'     \item{\emph{Library size parameters}}{
 #'         \describe{
 #'             \item{\code{lib.loc}}{Location (meanlog) parameter for the
-#'             library size log-normal distribution.}
+#'             library size log-normal distribution, or mean parameter if a
+#'             normal distribution is used.}
 #'             \item{\code{lib.scale}}{Scale (sdlog) parameter for the library
-#'             size log-normal distribution.}
+#'             size log-normal distribution, or sd parameter if a normal
+#'             distribution is used.}
+#'             \item{\code{lib.norm}}{Logical. Whether to use a normal
+#'             distribution for library sizes instead of a log-normal.}
 #'         }
 #'     }
 #'     \item{\emph{Expression outlier parameters}}{
@@ -142,8 +146,12 @@ setClass("SimpleParams",
 #'     }
 #'     \item{\emph{Dropout parameters}}{
 #'         \describe{
-#'             \item{\code{dropout.present}}{Logical. Whether to simulate
+#'             \item{\code{dropout.type}}{The type of dropout to simulate.
-#'             dropout.}
+#'             "none" indicates no dropout, "experiment" is global dropout using
+#'             the same parameters for every cell, "batch" uses the same
+#'             parameters for every cell in each batch, "group" uses the same
+#'             parameters for every cell in each groups and "cell" uses a
+#'             different set of parameters for each cell.}
 #'             \item{\code{dropout.mid}}{Midpoint parameter for the dropout
 #'             logistic function.}
 #'             \item{\code{dropout.shape}}{Shape parameter for the dropout
@@ -197,6 +205,7 @@ setClass("SplatParams",
                   mean.rate = "numeric",
                   lib.loc = "numeric",
                   lib.scale = "numeric",
+                   lib.norm = "logical",
                   out.prob = "numeric",
                   out.facLoc = "numeric",
                   out.facScale = "numeric",
@@ -208,7 +217,7 @@ setClass("SplatParams",
                   de.facScale = "numeric",
                   bcv.common = "numeric",
                   bcv.df = "numeric",
-                   dropout.present = "logical",
+                   dropout.type = "character",
                   dropout.mid = "numeric",
                   dropout.shape = "numeric",
                   path.from = "numeric",
@@ -224,6 +233,7 @@ setClass("SplatParams",
                               mean.shape = 0.6,
                               lib.loc = 11,
                               lib.scale = 0.2,
+                               lib.norm = FALSE,
                               out.prob = 0.05,
                               out.facLoc = 4,
                               out.facScale = 0.5,
@@ -235,7 +245,7 @@ setClass("SplatParams",
                               de.facScale = 0.4,
                               bcv.common = 0.1,
                               bcv.df = 60,
-                               dropout.present = FALSE,
+                               dropout.type = "none",
                               dropout.mid = 0,
                               dropout.shape = -1,
                               path.from = 0,
@@ -657,3 +667,58 @@ setClass("ZINBParams",
         contains = "Params",
         slots = c(model = "ANY"),
         prototype = prototype(nGenes = 100, nCells = 50))
+#' The SparseDCParams class
+#'
+#' S4 class that holds parameters for the SparseDC simulation.
+#'
+#' @section Parameters:
+#'
+#' The SparseDC simulation uses the following parameters:
+#'
+#' \describe{
+#'     \item{\code{nGenes}}{The number of genes to simulate in each condition.}
+#'     \item{\code{nCells}}{The number of cells to simulate.}
+#'     \item{\code{[seed]}}{Seed to use for generating random numbers.}
+#'     \item{\code{markers.n}}{Number of marker genes to simulate for each
+#'     cluster.}
+#'     \item{\code{markers.shared}}{Number of marker genes for each cluster
+#'     shared between conditions. Must be less than or equal to
+#'     \code{markers.n}}.
+#'     \item{\code{[markers.same]}}{Logical. Whether each cluster should have
+#'     the same set of marker genes.}
+#'     \item{\code{clusts.c1}}{Numeric vector of clusters present in
+#'     condition 1. The number of times a cluster is repeated controls the
+#'     proportion of cells from that cluster.}
+#'     \item{\code{clusts.c2}}{Numeric vector of clusters present in
+#'     condition 2. The number of times a cluster is repeated controls the
+#'     proportion of cells from that cluster.}
+#'     \item{\code{[mean.lower]}}{Lower bound for cluster gene means.}
+#'     \item{\code{[mean.upper]}}{Upper bound for cluster gene means.}
+#' }
+#'
+#' The parameters not shown in brackets can be estimated from real data using
+#' \code{\link{sparseDCEstimate}}. For details of the SparseDC simulation
+#' see \code{\link{sparseDCSimulate}}.
+#'
+#' @name SparseDCParams
+#' @rdname SparseDCParams
+#' @aliases SparseDCParams-class
+#' @exportClass SparseDCParams
+setClass("SparseDCParams",
+         contains = "Params",
+         slots = c(markers.n = "numeric",
+                   markers.shared = "numeric",
+                   markers.same = "logical",
+                   clusts.c1 = "numeric",
+                   clusts.c2 = "numeric",
+                   mean.lower = "numeric",
+                   mean.upper = "numeric"),
+         prototype = prototype(markers.n = 0,
+                               markers.shared = 0,
+                               markers.same = FALSE,
+                               clusts.c1 = 1,
+                               clusts.c2 = 1,
+                               mean.lower = 1,
+                               mean.upper = 2))
--- a/R/BASiCS-estimate.R
+++ b/R/BASiCS-estimate.R
@@ -2,8 +2,8 @@
 #'
 #' Estimate simulation parameters for the BASiCS simulation from a real dataset.
 #'
-#' @param counts either a counts matrix or an SCESet object containing count
+#' @param counts either a counts matrix or a SingleCellExperiment object
-#'        data to estimate parameters from.
+#'        containing count data to estimate parameters from.
 #' @param spike.info data.frame describing spike-ins with two columns: "Name"
 #'        giving the names of the spike-in features (must match
 #'        \code{rownames(counts)}) and "Input" giving the number of input
@@ -31,7 +31,10 @@
 #'
 #' @examples
 #' \dontrun{
+#' # Load example data
+#' library(scater)
 #' data("sc_example_counts")
+#'
 #' spike.info <- data.frame(Name = rownames(sc_example_counts)[1:10],
 #'                          Input = rnorm(10, 500, 200),
 #'                          stringsAsFactors = FALSE)
@@ -49,10 +52,12 @@ BASiCSEstimate <- function(counts, spike.info = NULL, batch = NULL,
 #' @rdname BASiCSEstimate
 #' @export
-BASiCSEstimate.SCESet <- function(counts, spike.info = NULL, batch = NULL,
+BASiCSEstimate.SingleCellExperiment <- function(counts, spike.info = NULL,
-                                  n = 20000, thin = 10, burn = 5000,
+                                                batch = NULL, n = 20000,
-                                  params = newBASiCSParams(), verbose = TRUE,
+                                                thin = 10, burn = 5000,
-                                  progress = TRUE, ...) {
+                                                params = newBASiCSParams(),
+                                                verbose = TRUE, progress = TRUE,
+                                                ...) {
    counts <- BiocGenerics::counts(counts)
    BASiCSEstimate(counts, params)
 }

--- a/R/BASiCS-simulate.R
+++ b/R/BASiCS-simulate.R
@@ -10,8 +10,8 @@
 #' @details
 #' This function is just a wrapper around \code{\link[BASiCS]{BASiCS_Sim}} that
 #' takes a \code{\link{BASiCSParams}}, runs the simulation then converts the
-#' output to an \code{\link[scater]{SCESet}} object. See
+#' output to a \code{\link[SingleCellExperiment]{SingleCellExperiment}} object.
-#' \code{\link[BASiCS]{BASiCS_Sim}} for more details of how the simulation
+#' See \code{\link[BASiCS]{BASiCS_Sim}} for more details of how the simulation
 #' works.
 #'
 #' @return SingleCellExperiment containing simulated counts
@@ -131,7 +131,7 @@ BASiCSSimulate <- function(params = newBASiCSParams(), verbose = TRUE, ...) {
    sim <- SingleCellExperiment(assays = list(counts = counts),
                                rowData = features,
                                colData = cells,
-                                metadata = list(params = params))
+                                metadata = list(Params = params))
    if (verbose) {message("Done!")}
    return(sim)

--- a/R/BASiCSParams-methods.R
+++ b/R/BASiCSParams-methods.R
@@ -68,7 +68,7 @@ setValidity("BASiCSParams", function(object) {
 })
 #' @rdname setParam
-setMethod("setParam", "BASiCSParams",function(object, name, value) {
+setMethod("setParam", "BASiCSParams", function(object, name, value) {
    checkmate::assertString(name)
    if (name == "nCells" || name == "nBatches") {
@@ -87,7 +87,9 @@ setMethod("setParam", "BASiCSParams",function(object, name, value) {
 setMethod("show", "BASiCSParams", function(object) {
-    pp <- list("Batches:"     = c("(Batches)"     = "nBatches",
+    pp <- list("Genes:"       = c("(Params)"      = "gene.params"),
+               "Cells:"       = c("(Params)"      = "cell.params"),
+               "Batches:"     = c("(Batches)"     = "nBatches",
                                  "(Batch Cells)" = "batchCells"),
               "Spike-ins:"   = c("(Number)"      = "nSpikes",
                                  "(Means)"       = "spike.means"),
@@ -95,20 +97,6 @@ setMethod("show", "BASiCSParams", function(object) {
    callNextMethod()
-    gene.params <- getParam(object, "gene.params")
-    cell.params <- getParam(object, "cell.params")
-    cat("Genes:", "\n")
-    cat("(Params)", "\n")
-    cat("data.frame with", dim(gene.params)[1], "features\n")
-    print(head(gene.params, n = 3))
-    cat("  ...  ...\n\n")
-    cat("Cells:", "\n")
-    cat("(Params)", "\n")
-    cat("data.frame with", dim(cell.params)[1], "features\n")
-    print(head(cell.params, n = 3))
-    cat("  ...  ...\n\n")
    showPP(object, pp)
 })

--- a/R/Lun2Params-methods.R
+++ b/R/Lun2Params-methods.R
@@ -81,7 +81,9 @@ setMethod("setParam", "Lun2Params", function(object, name, value) {
 setMethod("show", "Lun2Params", function(object) {
-    pp <- list("Plates:"    = c("[Number]"        = "nPlates",
+    pp <- list("Genes:"     = c("(Params)"        = "gene.params",
+                                "(ZI Params)"     = "zi.params"),
+               "Plates:"    = c("[Number]"        = "nPlates",
                                "[Modifier]"      = "plate.mod",
                                "(Variance)"      = "plate.var"),
               "Cells:"     = c("[Plates]"        = "cell.plates",
@@ -92,17 +94,5 @@ setMethod("show", "Lun2Params", function(object) {
    callNextMethod()
-    gene.params <- getParam(object, "gene.params")
-    zi.params <- getParam(object, "zi.params")
-    cat("Genes:", "\n")
-    cat("(Params)", "\n")
-    cat("data.frame with", dim(gene.params)[1], "features\n")
-    print(head(gene.params, n = 3))
-    cat("  ...  ...\n")
-    cat("(ZI Params)", "\n")
-    cat("data.frame with", dim(zi.params)[1], "features\n")
-    print(head(zi.params, n = 3))
-    cat("  ...  ...  ...\n\n")
    showPP(object, pp)
 })
--- a/R/Params-methods.R
+++ b/R/Params-methods.R
@@ -28,9 +28,10 @@ setMethod("show", "Params", function(object) {
                             "(Cells)" = "nCells",
                             "[Seed]"  = "seed"))
-    cat("A Params object of class", class(object), "\n")
+    cat("A", crayon::bold("Params"), "object of class",
-    cat("Parameters can be (estimable) or [not estimable],",
+        crayon::bold(class(object)), "\n")
-        "'Default' or 'NOT DEFAULT'.", "\n\n")
+    cat("Parameters can be (estimable) or", crayon::blue("[not estimable],"),
+        "'Default' or ", crayon::bold(crayon::green("'NOT DEFAULT'")), "\n\n")
    showPP(object, pp)
    cat(length(slotNames(object)) - 3, "additional parameters", "\n\n")
 })
@@ -48,4 +49,4 @@ setMethod("expandParams", "Params", function(object, vectors, n) {
    object <- setParamsUnchecked(object, update)
    return(object)
 })
\ No newline at end of file
--- a/R/SCE-functions.R
+++ b/R/SCE-functions.R
@@ -14,10 +14,10 @@
 #' @details
 #' Currently adds the following statistics: mean, variance, coefficient of
 #' variation, median and median absolute deviation. Statistics are added to
-#' the \code{\link[SummarizedExperiment]{rowData}} slot and are named
+#' the \code{\link{rowData}} slot and are named \code{Stat[Log]Value[No0]} where
-#' \code{Stat[Log]Value[No0]} where \code{Log} and \code{No0} are added if those
+#' \code{Log} and \code{No0} are added if those arguments are true.
-#' arguments are true. UpperCamelCase is used to differentiate these columns
+#' UpperCamelCase is used to differentiate these columns from those added by
-#' from those added by analysis packages.
+#' analysis packages.
 #'
 #' @return SingleCellExperiment with additional feature statistics
 #'
@@ -88,7 +88,7 @@ addFeatureStats <- function(sce, value = c("counts", "cpm", "tpm", "fpkm"),
 #'
 #' @details
 #' This function adds simulated gene lengths to the
-#' \code{\link[SummarizedExperiment]{rowData}} slot of a
+#' \code{\link{rowData}} slot of a
 #' \code{\link[SingleCellExperiment]{SingleCellExperiment}} object that can be
 #' used for calculating length normalised expression values such as TPM or FPKM.
 #' The \code{generate} method simulates lengths using a (rounded) log-normal

--- a/R/SparseDCParams-methods.R
+++ b/R/SparseDCParams-methods.R
+#' @rdname newParams
+#' @importFrom methods new
+#' @export
+newSparseDCParams <- function(...) {
+    if (!requireNamespace("SparseDC", quietly = TRUE)) {
+        stop("The SparseDC simulation requires the 'SparseDC' package.")
+    }
+    params <- new("SparseDCParams")
+    params <- setParams(params, ...)
+    return(params)
+}
+setValidity("SparseDCParams", function(object) {
+    v <- getParams(object, slotNames(object))
+    checks <- c(nGenes = checkmate::checkInt(v$nGenes, lower = 1),
+                nCells = checkmate::checkInt(v$nCells, lower = 1),
+                markers.n = checkmate::checkInt(v$markers.n, lower = 0,
+                                                upper = v$nGenes),
+                markers.shared = checkmate::checkInt(v$markers.shared,
+                                                     lower = 0,
+                                                     upper = v$markers.n),
+                markers.same = checkmate::checkFlag(v$markers.same),
+                clusts.c1 = checkmate::checkIntegerish(v$clusts.c1, lower = 1,
+                                                       any.missing = FALSE,
+                                                       min.len = 1),
+                clusts.c2 = checkmate::checkIntegerish(v$clusts.c2, lower = 1,
+                                                       any.missing = FALSE,
+                                                       min.len = 1),
+                mean.lower = checkmate::checkNumber(v$mean.lower,
+                                                    finite = TRUE),
+                mean.upper = checkmate::checkNumber(v$mean.upper,
+                                                    finite = TRUE),
+                seed = checkmate::checkInt(v$seed, lower = 0))
+    if (length(v$clusts.c1) == 1) {
+        if (v$clusts.c1 > 1) {
+            checks <- c(checks,
+                        clusts.c1 = "If clusts.c1 is length 1 it must equal 1")
+        }
+    }
+    if (length(v$clusts.c2) == 1) {
+        if (v$clusts.c2 > 1) {
+            checks <- c(checks,
+                        clusts.c2 = "If clusts.c2 is length 1 it must equal 1")
+        }
+    }
+    clusts.combined <- c(v$clusts.c1, v$clusts.c2)
+    if (!all(seq_len(max(clusts.combined)) %in% clusts.combined)) {
+        checks <- c(checks, clusts = "Cluster labels must be sequential")
+    }
+    if (all(checks == TRUE)) {
+        valid <- TRUE
+    } else {
+        valid <- checks[checks != TRUE]
+        valid <- paste(names(valid), valid, sep = ": ")
+    }
+    return(valid)
+})
+setMethod("show", "SparseDCParams", function(object) {
+    pp <- list("Markers:"  = c("(Number)"      = "markers.n",
+                               "(Shared)"      = "markers.shared",
+                               "[Same]"        = "markers.same"),
+               "Clusters:" = c("(Condition 1)" = "clusts.c1",
+                               "(Condition 2)" = "clusts.c2"),
+               "Means:"    = c("[Lower]"       = "mean.lower",
+                               "[Upper]"       = "mean.upper"))
+    callNextMethod()
+    showPP(object, pp)
+})
--- a/R/SplatParams-methods.R
+++ b/R/SplatParams-methods.R
@@ -30,6 +30,7 @@ setValidity("SplatParams", function(object) {
                mean.shape = checkNumber(v$mean.shape, lower = 0),
                lib.loc = checkNumber(v$lib.loc),
                lib.scale = checkNumber(v$lib.scale, lower = 0),
+                lib.norm = checkFlag(v$lib.norm),
                out.prob = checkNumber(v$out.prob, lower = 0, upper = 1),
                out.facLoc = checkNumber(v$out.facLoc),
                out.facScale = checkNumber(v$out.facScale, lower = 0),
@@ -45,9 +46,12 @@ setValidity("SplatParams", function(object) {
                                           len = nGroups),
                bcv.common = checkNumber(v$bcv.common, lower = 0),
                bcv.df = checkNumber(v$bcv.df, lower = 0),
-                dropout.present = checkFlag(v$dropout.present),
+                dropout.type = checkCharacter(v$dropout.type, len = 1,
-                dropout.mid = checkNumber(v$dropout.mid),
+                                              any.missing = FALSE),
-                dropout.shape = checkNumber(v$dropout.shape),
+                dropout.mid = checkNumeric(v$dropout.mid, finite = TRUE,
+                                           any.missing = FALSE, min.len = 1),
+                dropout.shape = checkNumeric(v$dropout.shape, finite = TRUE,
+                                             any.missing = FALSE, min.len = 1),
                path.from = checkIntegerish(v$path.from, lower = 0,
                                            upper = nGroups, len = nGroups),
                path.length = checkIntegerish(v$path.length, lower = 1,
@@ -74,7 +78,15 @@ setValidity("SplatParams", function(object) {
    if (!(0 %in% v$path.from)) {
       checks <- c(checks, path.from = "origin must be specified in path.from")
    } else if (any(v$path.from == seq_len(nGroups))) {
-        checks <- c(checks, stop("path cannot begin at itself"))
+        checks <- c(checks, "path cannot begin at itself")
+    }
+    # Check dropout type
+    if (!(v$dropout.type %in%
+          c("none", "experiment", "batch", "group", "cell"))) {
+        checks <- c(checks,
+                    paste("dropout.type must be one of: 'none', 'experiment',",
+                          "'batch', 'group', 'cell'"))
    }
    if (all(checks == TRUE)) {
@@ -106,6 +118,50 @@ setMethod("setParam", "SplatParams",function(object, name, value) {
    if (name == "group.prob") {
        object <- setParamUnchecked(object, "nGroups", length(value))
+        path.from <- getParam(object, "path.from")
+        if (length(path.from) > 1 & length(path.from) != length(value)) {
+            warning("nGroups has changed, resetting path.from")
+            object <- setParam(object, "path.from", 0)
+        }
+    }
+    if (name == "dropout.type") {
+        mid.len <- length(getParam(object, "dropout.mid"))
+        mid.shape <- length(getParam(object, "dropout.shape"))
+        if ((value == "experiment")) {
+            if ((mid.len != 1) | (mid.shape != 1)) {
+                stop("dropout.type cannot be set to 'experiment' because ",
+                     "dropout.mid and dropout.shape aren't length 1, ",
+                     "set dropout.mid and dropout.shape first")
+            }
+        }
+        if ((value == "batch")) {
+            n <- getParam(object, "nBatches")
+            if ((mid.len != n) | (mid.shape != n)) {
+                stop("dropout.type cannot be set to 'batch' because ",
+                     "dropout.mid and dropout.shape aren't length equal to ",
+                     "nBatches (", n, "), set dropout.mid and dropout.shape ",
+                     "first")
+            }
+        }
+        if ((value == "group")) {
+            n <- getParam(object, "nGroups")
+            if ((mid.len != n) | (mid.shape != n)) {
+                stop("dropout.type cannot be set to 'group' because ",
+                     "dropout.mid and dropout.shape aren't length equal to ",
+                     "nGroups (", n, "), set dropout.mid and dropout.shape ",
+                     "first")
+            }
+        }
+        if ((value == "cell")) {
+            n <- getParam(object, "nCells")
+            if ((mid.len != n) | (mid.shape != n)) {
+                stop("dropout.type cannot be set to 'cell' because ",
+                     "dropout.mid and dropout.shape aren't length equal to ",
+                     "nCells (", n, "), set dropout.mid and dropout.shape ",
+                     "first")
+            }
+        }
    }
    object <- callNextMethod()
@@ -123,7 +179,8 @@ setMethod("show", "SplatParams", function(object) {
               "Mean:"           = c("(Rate)"         = "mean.rate",
                                     "(Shape)"        = "mean.shape"),
               "Library size:"   = c("(Location)"     = "lib.loc",
-                                     "(Scale)"        = "lib.scale"),
+                                     "(Scale)"        = "lib.scale",
+                                     "(Norm)"         = "lib.norm"),
               "Exprs outliers:" = c("(Probability)"  = "out.prob",
                                     "(Location)"     = "out.facLoc",
                                     "(Scale)"        = "out.facScale"),
@@ -135,7 +192,7 @@ setMethod("show", "SplatParams", function(object) {
                                     "[Scale]"        = "de.facScale"),
               "BCV:"            = c("(Common Disp)"  = "bcv.common",
                                     "(DoF)"          = "bcv.df"),
-               "Dropout:"        = c("[Present]"      = "dropout.present",
+               "Dropout:"        = c("[Type]"         = "dropout.type",
                                     "(Midpoint)"     = "dropout.mid",
                                     "(Shape)"        = "dropout.shape"),
               "Paths:"          = c("[From]"         = "path.from",

--- a/R/ZINBParams-methods.R
+++ b/R/ZINBParams-methods.R
@@ -90,12 +90,20 @@ setMethod("show", "ZINBParams", function(object) {
    callNextMethod()
    model <- getParam(object, "model")
-    cat("Model:", "\n")
-    cat("ZinbModel with", zinbwave::nFeatures(model), "features,",
-        zinbwave::nSamples(model), "samples,", zinbwave::nFactors(model),
-        "latent factors and", zinbwave::nParams(model), "parameters", "\n\n")
    default <- zinbwave::zinbModel()
+    not.default <- !identical(model, default)
+    cat(crayon::bold("Model:"), "\n")
+    msg <- paste("ZinbModel with", zinbwave::nFeatures(model), "features,",
+                 zinbwave::nSamples(model), "samples,",
+                 zinbwave::nFactors(model), "latent factors and",
+                 zinbwave::nParams(model), "parameters")
+    if (not.default) {
+        msg <- crayon::bold(crayon::green(msg))
+    }
+    cat(msg, "\n\n")
+    cat(crayon::bold("Parameters of the ZinbModel"), "\n\n")
    for (category in names(pp)) {
        parameters <- pp[[category]]
        values <- lapply(parameters, function(x) {slot(model, x)})
@@ -118,9 +126,8 @@ setMethod("show", "ZINBParams", function(object) {
        default.values <- lapply(parameters, function(x) {slot(default, x)})
        default.values <- sapply(default.values, paste, collapse = ", ")
        not.default <- values != default.values
-        names(short.values)[not.default] <- toupper(names(values[not.default]))
+        cat(crayon::bold(c("Model", category)), "\n")
-        cat("Model", category, "\n")
+        showValues(short.values, not.default)
-        print(noquote(short.values), print.gap = 2)
        cat("\n")
    }

--- a/R/compare.R
+++ b/R/compare.R
@@ -79,8 +79,8 @@ compareSCEs <- function(sces, point.size = 0.1, point.alpha = 0.1,
        sce <- addFeatureStats(sce, "counts")
        sce <- addFeatureStats(sce, "cpm")
        sce <- addFeatureStats(sce, "cpm", log = TRUE)
-        colData(sce)$PctZero <- 100 * (1 - colData(sce)$total_features /
+        n.features <- colData(sce)$total_features_by_counts
-                                           nrow(sce))
+        colData(sce)$PctZero <- 100 * (1 - n.features / nrow(sce))
        sces[[name]] <- sce
    }
@@ -140,7 +140,7 @@ compareSCEs <- function(sces, point.size = 0.1, point.alpha = 0.1,
        theme_minimal()
    z.gene <- ggplot(features,
-                     aes_string(x = "Dataset", y = "pct_dropout_counts",
+                     aes_string(x = "Dataset", y = "pct_dropout_by_counts",
                                colour = "Dataset")) +
        geom_boxplot() +
        scale_y_continuous(limits = c(0, 100)) +
@@ -160,7 +160,8 @@ compareSCEs <- function(sces, point.size = 0.1, point.alpha = 0.1,
        theme_minimal()
    mean.zeros <- ggplot(features,
-                         aes_string(x = "MeanCounts", y = "pct_dropout_counts",
+                         aes_string(x = "MeanCounts",
+                                    y = "pct_dropout_by_counts",
                                    colour = "Dataset", fill = "Dataset")) +
        geom_point(size = point.size, alpha = point.alpha) +
        scale_x_log10(labels = scales::comma) +
@@ -296,7 +297,7 @@ diffSCEs <- function(sces, ref, point.size = 0.1, point.alpha = 0.1,
    for (name in names(sces)) {
        sce <- sces[[name]]
        if (!identical(dim(sce), ref.dim)) {
-            stop("SCESets must have the same dimensions")
+            stop("all datasets in 'sces' must have the same dimensions")
        }
        rowData(sce)$Dataset <- name
        colData(sce)$Dataset <- name
@@ -304,8 +305,9 @@ diffSCEs <- function(sces, ref, point.size = 0.1, point.alpha = 0.1,
        cpm(sce) <- scater::calculateCPM(sce, use_size_factors = FALSE)
        sce <- addFeatureStats(sce, "counts")
        sce <- addFeatureStats(sce, "cpm", log = TRUE)
-        colData(sce)$PctZero <- 100 * (1 - colData(sce)$total_features /
+        n.features <- colData(sce)$total_features_by_counts
-                                                              nrow(sce))
+        colData(sce)$PctZero <- 100 * (1 - n.features / nrow(sce))
+        rowData(sce)$RankCounts <- rank(rowData(sce)$mean_counts)
        sces[[name]] <- sce
    }
@@ -314,12 +316,12 @@ diffSCEs <- function(sces, ref, point.size = 0.1, point.alpha = 0.1,
    ref.means <- sort(rowData(ref.sce)$MeanLogCPM)
    ref.vars <- sort(rowData(ref.sce)$VarLogCPM)
    ref.libs <- sort(colData(ref.sce)$total_counts)
-    ref.z.gene <- sort(rowData(ref.sce)$pct_dropout_counts)
+    ref.z.gene <- sort(rowData(ref.sce)$pct_dropout_by_counts)
    ref.z.cell <- sort(colData(ref.sce)$PctZero)
-    ref.rank.ord <- order(rowData(ref.sce)$rank_counts)
+    ref.rank.ord <- order(rowData(ref.sce)$RankCounts)
    ref.vars.rank <- rowData(ref.sce)$VarLogCPM[ref.rank.ord]
-    ref.z.gene.rank <- rowData(ref.sce)$pct_dropout_counts[ref.rank.ord]
+    ref.z.gene.rank <- rowData(ref.sce)$pct_dropout_by_counts[ref.rank.ord]
    for (name in names(sces)) {
        sce <- sces[[name]]
@@ -334,8 +336,8 @@ diffSCEs <- function(sces, ref, point.size = 0.1, point.alpha = 0.1,
        colData(sce)$RankDiffLibSize <- colData(sce)$total_counts -
            colData(sce)$RefRankLibSize
        rowData(sce)$RefRankZeros <- ref.z.gene[rank(
-                                               rowData(sce)$pct_dropout_counts)]
+                                               rowData(sce)$pct_dropout_by_counts)]
-        rowData(sce)$RankDiffZeros <- rowData(sce)$pct_dropout_counts -
+        rowData(sce)$RankDiffZeros <- rowData(sce)$pct_dropout_by_counts -
            rowData(sce)$RefRankZeros
        colData(sce)$RefRankZeros <- ref.z.cell[rank(
                                               colData(sce)$PctZero)]
@@ -343,9 +345,9 @@ diffSCEs <- function(sces, ref, point.size = 0.1, point.alpha = 0.1,
            colData(sce)$RefRankZeros
        rowData(sce)$MeanRankVarDiff <- rowData(sce)$VarLogCPM -
-            ref.vars.rank[rowData(sce)$rank_counts]
+            ref.vars.rank[rowData(sce)$RankCounts]
-        rowData(sce)$MeanRankZerosDiff <- rowData(sce)$pct_dropout_counts -
+        rowData(sce)$MeanRankZerosDiff <- rowData(sce)$pct_dropout_by_counts -
-            ref.z.gene.rank[rowData(sce)$rank_counts]
+            ref.z.gene.rank[rowData(sce)$RankCounts]
        sces[[name]] <- sce
    }
@@ -391,7 +393,7 @@ diffSCEs <- function(sces, ref, point.size = 0.1, point.alpha = 0.1,
        theme_minimal()
    mean.var <- ggplot(features,
-                       aes_string(x = "rank_counts", y = "MeanRankVarDiff",
+                       aes_string(x = "RankCounts", y = "MeanRankVarDiff",
                                  colour = "Dataset", fill = "Dataset")) +
        geom_hline(yintercept = 0, colour = "red") +
        geom_point(size = point.size, alpha = point.alpha) +
@@ -434,7 +436,7 @@ diffSCEs <- function(sces, ref, point.size = 0.1, point.alpha = 0.1,
        theme_minimal()
    mean.zeros <- ggplot(features,
-                       aes_string(x = "rank_counts", y = "MeanRankZerosDiff",
+                       aes_string(x = "RankCounts", y = "MeanRankZerosDiff",
                                  colour = "Dataset", fill = "Dataset")) +
        geom_hline(yintercept = 0, colour = "red") +
        geom_point(size = point.size, alpha = point.alpha) +
@@ -479,7 +481,8 @@ diffSCEs <- function(sces, ref, point.size = 0.1, point.alpha = 0.1,
        theme_minimal()
    z.gene.qq <- ggplot(features,
-                        aes_string(x = "RefRankZeros", y = "pct_dropout_counts",
+                        aes_string(x = "RefRankZeros",
+                                   y = "pct_dropout_by_counts",
                                   colour = "Dataset")) +
        geom_abline(intercept = 0, slope = 1, colour = "red") +
        geom_point(size = point.size, alpha = point.alpha) +
@@ -714,8 +717,8 @@ makeDiffPanel <- function(diff, title = "Difference comparison",
 #' \dontrun{
 #' sim1 <- splatSimulate(nGenes = 1000, batchCells = 20)
 #' sim2 <- simpleSimulate(nGenes = 1000, nCells = 20)
-#' comparison <- compSCESets(list(Splat = sim1, Simple = sim2))
+#' comparison <- compSCEs(list(Splat = sim1, Simple = sim2))
-#' difference <- diffSCESets(list(Splat = sim1, Simple = sim2), ref = "Simple")
+#' difference <- diffSCEs(list(Splat = sim1, Simple = sim2), ref = "Simple")
 #' panel <- makeOverallPanel(comparison, difference)
 #' }
 #'

--- a/R/listSims.R
+++ b/R/listSims.R
@@ -71,7 +71,12 @@ listSims <- function(print = TRUE) {
                   "drisso/zinbwave",
                   "The ZINB-WaVE simulation simulates counts from a
                   sophisticated zero-inflated negative-binomial distribution
-                   including cell and gene-level covariates."))
+                   including cell and gene-level covariates."),
+                 c("SparseDC", "sparseDC", "10.1093/nar/gkx1113",
+                   "cran/SparseDC",
+                   "The SparseDC simulation simulates a set of clusters
+                   across two conditions, where some clusters may be present in
+                   only one condition."))
    sims.table <- data.frame(Name        = rep(NA, length(sims)),
                             Prefix      = rep(NA, length(sims)),
@@ -89,8 +94,10 @@ listSims <- function(print = TRUE) {
        cat("Splatter currently contains", length(sims), "simulations", "\n\n")
        for (idx in seq_len(nrow(sims.table))) {
            sim <- as.character(sims.table[idx, ])
-            cat(sim[1], paste0("(", sim[2], ")"), "\n")
+            cat(crayon::bold(sim[1]), crayon::yellow(paste0("(", sim[2], ")")),
-            cat("DOI:", sim[3], "\t", "GitHub:", sim[4], "\n")
+                "\n")
+            cat(crayon::bold("DOI:"), crayon::cyan(sim[3]), "\t",
+                crayon::bold("GitHub:"), crayon::cyan(sim[4]), "\n")
            cat(sim[5], "\n\n")
        }
    }

--- a/R/lun-estimate.R
+++ b/R/lun-estimate.R
@@ -14,7 +14,10 @@
 #' @return LunParams object containing the estimated parameters.
 #'
 #' @examples
+#' # Load example data
+#' library(scater)
 #' data("sc_example_counts")
+#'
 #' params <- lunEstimate(sc_example_counts)
 #' params
 #' @export

--- a/R/lun-simulate.R
+++ b/R/lun-simulate.R
@@ -132,7 +132,7 @@ lunSimulate <- function(params = newLunParams(), verbose = TRUE, ...) {
                                              CellMeans = cell.means),
                                rowData = features,
                                colData = cells,
-                                metadata = list(params = params))
+                                metadata = list(Params = params))
    if (verbose) {message("Done!")}

--- a/R/lun2-estimate.R
+++ b/R/lun2-estimate.R
@@ -9,9 +9,9 @@
 #' @param min.size minimum size of clusters when identifying group of cells in
 #'        the data.
 #' @param verbose logical. Whether to show progress messages.
-#' @param BPPARAM A \code{\link[BiocParallel]{BiocParallelParam}} instance
+#' @param BPPARAM A \code{\link{BiocParallelParam}} instance giving the parallel
-#'        giving the parallel back-end to be used. Default is
+#'        back-end to be used. Default is \code{\link{SerialParam}} which uses a
-#'        \code{\link[BiocParallel]{SerialParam}} which uses a single core.
+#'        single core.
 #'
 #' @details
 #' See \code{\link{Lun2Params}} for more details on the parameters.
@@ -20,8 +20,11 @@
 #'
 #' @examples
 #' \dontrun{
+#' # Load example data
+#' library(scater)
 #' data("sc_example_counts")
 #' data("sc_example_cell_info")
+#'
 #' plates <- factor(sc_example_cell_info$Mutation_Status)
 #' params <- lun2Estimate(sc_example_counts, plates, min.size = 20)
 #' params

--- a/R/lun2-simulate.R
+++ b/R/lun2-simulate.R
@@ -183,7 +183,7 @@ lun2Simulate <- function(params = newLun2Params(), zinb = FALSE,
                                              TrueCounts <- true.counts),
                                rowData = features,
                                colData = cells,
-                                metadata = list(params = params))
+                                metadata = list(Params = params))
    if (zinb) {
        rownames(is.zero) <- gene.names

--- a/R/mfa-estimate.R
+++ b/R/mfa-estimate.R
@@ -15,7 +15,10 @@
 #' @return MFAParams object containing the estimated parameters.
 #'
 #' @examples
+#' # Load example data
+#' library(scater)
 #' data("sc_example_counts")
+#'
 #' params <- mfaEstimate(sc_example_counts)
 #' params
 #' @export