@@ -1,11 +1,11 @@

 Package: vsclust

 Type: Package

 Title: Feature-based variance-sensitive quantitative clustering

-Version: 0.99.20

+Version: 0.99.21

 Date: 2022-03-23

 Authors@R: 

     person(

-        "Veit", "Schwämmle",

+        "Veit", "Schwaemmle",

         email = "[email protected]",

         role = c("aut", "cre")

     )

@@ -26,7 +26,8 @@ Suggests:

     yaml,

     testthat (>= 3.0.0),

     rmarkdown,

-    BiocStyle

+    BiocStyle,

+    clusterProfiler

 LinkingTo: Rcpp

 biocViews: 

   Clustering,

@@ -18,12 +18,15 @@ export(pcaWithVar)

 export(runClustWrapper)

 export(runVSClustApp)

 export(vsclust_algorithm)

-import(MultiAssayExperiment)

 import(grDevices)

 import(graphics)

 import(limma)

 import(parallel)

 import(stats)

+importFrom(MultiAssayExperiment,assay)

+importFrom(MultiAssayExperiment,assays)

+importFrom(MultiAssayExperiment,colData)

+importFrom(MultiAssayExperiment,sampleMap)

 importFrom(grDevices,rainbow)

 importFrom(grDevices,recordPlot)

 importFrom(matrixStats,rowMaxs)

@@ -31,7 +31,7 @@ cvalidate.xiebeni <-

     ncenters <-

       dim(clres$centers)[1]

     for (i in seq_len(ncenters - 1)) {

-      for (j in (i + 1):ncenters) {

+      for (j in seq(i + 1,ncenters, 1)) {

         diff <- clres$ce[i,] - clres$ce[j,]

         diffdist <-

           t(diff) %*% t(t(diff))

@@ -61,9 +61,9 @@ cvalidate.xiebeni <-

 #' @param mfrow vector of two numbers for the number of rows and colums, figure 

 #' panels are distributed in the plot

 #' @param colo color map to be used (can be missing)

-#' @param min.mem filter for showing only features with a higher membership 

+#' @param minMem filter for showing only features with a higher membership 

 #' values than this value

-#' @param time.labels alternative labels for different conditions

+#' @param timeLabels alternative labels for different conditions

 #' @param filename for writing into pdf. Will write on screen when using NA

 #' @param xlab Label of x-axis

 #' @param ylab Label of y-axis

@@ -72,9 +72,9 @@ cvalidate.xiebeni <-

 #' data <- matrix(rnorm(seq_len(5000)), nrow=500)

 #' # Run clustering

 #' clres <- vsclust_algorithm(data, centers=2, m=1.5)

-#' mfuzz.plot(data, clres,  mfrow=c(2,3), min.mem=0.0)

+#' mfuzz.plot(data, clres,  mfrow=c(2,3), minMem=0.0)

 #' @return Multiple panels showing expression profiles of clustered features 

-#' passing the min.mem threshold

+#' passing the minMem threshold

 #' @export

 #' @references

 #' Schwaemmle V, Jensen ON. VSClust: feature-based variance-sensitive clustering 

@@ -95,15 +95,15 @@ mfuzz.plot <-

             cl,

             mfrow = c(1, 1),

             colo,

-            min.mem = 0,

-            time.labels,

+            minMem = 0,

+            timeLabels,

             filename = NA,

             xlab = "Time",

             ylab = "Expression changes")

   {

     clusterindex <- cl[[3]]

     memship <- cl[[4]]

-    memship[memship < min.mem] <- -1

+    memship[memship < minMem] <- -1

     colorindex <- integer(dim(dat)[[1]])

     if (missing(colo)) {

       colo <- c(

@@ -168,15 +168,15 @@ mfuzz.plot <-

       )

     }    else {

       if (colo == "fancy") {

-        fancy.blue <- c(c(255:0), rep(0, length(c(255:0))),

+        fancyBlue <- c(c(255:0), rep(0, length(c(255:0))),

                         rep(0, length(c(255:150))))

-        fancy.green <-

+        fanceGreen <-

           c(c(0:255), c(255:0), rep(0, length(c(255:150))))

-        fancy.red <- c(c(0:255), rep(255, length(c(255:0))),

+        fancyRed <- c(c(0:255), rep(255, length(c(255:0))),

                        c(255:150))

-        colo <- rgb(blue = fancy.blue / 255,

-                    green = fancy.green / 255,

-                    red = fancy.red / 255)

+        colo <- rgb(blue = fancyBlue / 255,

+                    green = fanceGreen / 255,

+                    red = fancyRed / 255)

       }

     }

     colorseq <- seq(0, 1, length = length(colo))

@@ -208,12 +208,12 @@ mfuzz.plot <-

             main = paste("Cluster", j),

             axes = FALSE

           )

-          if (missing(time.labels)) {

+          if (missing(timeLabels)) {

             axis(1, seq_len(dim(dat)[[2]]), c(seq_len(dim(dat)[[2]])))

             axis(2)

           }

           else {

-            axis(1, seq_len(dim(dat)[[2]]), time.labels)

+            axis(1, seq_len(dim(dat)[[2]]), timeLabels)

             axis(2)

           }

         }

@@ -235,12 +235,12 @@ mfuzz.plot <-

             main = paste("Cluster", j),

             axes = FALSE

           )

-          if (missing(time.labels)) {

+          if (missing(timeLabels)) {

             axis(1, seq_len(dim(dat)[[2]]), seq_len(dim(dat)[[2]]))

             axis(2)

           }

           else {

-            axis(1, seq_len(dim(dat)[[2]]), time.labels)

+            axis(1, seq_len(dim(dat)[[2]]), timeLabels)

             axis(2)

           }

         }

@@ -271,7 +271,7 @@ mfuzz.plot <-

 #'

 #' @param ClustInd Matrix with values from validity indices

 #' @return Multiple panels showing expression profiles of clustered features 

-#' passing the min.mem threshold

+#' passing the minMem threshold

 #' @examples

 #' data("artificial_clusters")

 #' dat <- averageCond(artificial_clusters, 5, 10)

@@ -297,7 +297,7 @@ estimClust.plot <- function(ClustInd) {

   par(mfrow = c(1, 3))

   maxClust <- nrow(ClustInd) + 2

   plot(

-    3:maxClust,

+    seq(3,maxClust,1),

     ClustInd[seq_len(nrow(ClustInd)), "MinCentroidDist_VSClust"],

     col = 2 ,

     type = "b",

@@ -309,7 +309,7 @@ estimClust.plot <- function(ClustInd) {

              max(ClustInd[, grep("MinCentroidDist",

                                  colnames(ClustInd))], na.rm = TRUE))

   )

-  lines(3:maxClust, ClustInd[seq_len(nrow(ClustInd)), "MinCentroidDist_FCM"],

+  lines(seq(3,maxClust,1), ClustInd[seq_len(nrow(ClustInd)), "MinCentroidDist_FCM"],

         col = 3, type = "b")

   dmindist <- optimalClustNum(ClustInd)

   points(dmindist,

@@ -325,7 +325,7 @@ estimClust.plot <- function(ClustInd) {

   )

   grid(NULL, NA, lwd = 1, col = 1)

   plot(

-    3:maxClust,

+    seq(3,maxClust,1),

     ClustInd[seq_len(nrow(ClustInd)), "XieBeni_VSClust"],

     col = 2,

     type = "b",

@@ -337,7 +337,7 @@ estimClust.plot <- function(ClustInd) {

              max(ClustInd[, grep("XieBeni", colnames(ClustInd))], na.rm =

                    TRUE))

   )

-  lines(3:maxClust, ClustInd[seq_len(nrow(ClustInd)), "XieBeni_FCM"], type =

+  lines(seq(3,maxClust,1), ClustInd[seq_len(nrow(ClustInd)), "XieBeni_FCM"], type =

           "b", col = 3)

   dxiebeni <- optimalClustNum(ClustInd, index = "XieBeni")

   points(dxiebeni,

@@ -43,7 +43,7 @@ determine_fuzz <- function(dims, NClust, Sds = 1) {

   ### d_i and d_t

   difunc <-

     function(c, D) {

-      x <- 0:c

+      x <- seq(0,c,1)

       sum(choose(c, x) / (x * D + 1) * (-1) ^ x)

     }

@@ -73,13 +73,13 @@ determine_fuzz <- function(dims, NClust, Sds = 1) {

 #' @param x a numeric data matrix

 #' @param centers Either numeric for number of clusters or numeric matrix with 

 #' center coordinates

-#' @param iter.max Numeric for maximum number of iterations

+#' @param iterMax Numeric for maximum number of iterations

 #' @param verbose Verbose information

 #' @param dist Distance to use for the calculation. We prefer "euclidean" 

 #' (default)

 #' @param m Fuzzifier value: numeric or vector of length equal to number of rows 

 #' of x

-#' @param rate.par (experimental) numeric value for punishing missing values

+#' @param ratePar (experimental) numeric value for punishing missing values

 #' @param weights numeric or vector of length equal to number of rows of x

 #' @param control list with arguments to vsclust algorithms (now only cutoff for 

 #' relative tolerance: reltol)

@@ -112,11 +112,11 @@ determine_fuzz <- function(dims, NClust, Sds = 1) {

 vsclust_algorithm <-

   function(x,

            centers,

-           iter.max = 100,

+           iterMax = 100,

            verbose = FALSE,

            dist = "euclidean",

            m = 2,

-           rate.par = NULL,

+           ratePar = NULL,

            weights = 1,

            control = list())

   {

@@ -157,11 +157,11 @@ vsclust_algorithm <-

     if (xcols != ncol(centers))

       stop("Must have same number of columns in 'x' and 'centers'.")

-    if (iter.max < 1)

-      stop("Argument 'iter.max' must be positive.")

+    if (iterMax < 1)

+      stop("Argument 'iterMax' must be positive.")

-    if (missing(rate.par)) {

-      rate.par <- 0

+    if (missing(ratePar)) {

+      ratePar <- 0

     }

     reltol <- control$reltol

@@ -194,14 +194,14 @@ vsclust_algorithm <-

                        weights,

                        m,

                        dist - 1,

-                       iter.max,

+                       iterMax,

                        reltol,

                        verbose,

                        u ,

                        1,

                        iter,

                        NA,

-                       rate.par)

+                       ratePar)

     # put modified values in retval

     retval <-

       list(

@@ -212,10 +212,10 @@ vsclust_algorithm <-

         ncenters = ncenters,

         m = m,

         dist = dist - 1,

-        iter.max = iter.max,

+        iterMax = iterMax,

         reltol = reltol,

         verbose = verbose,

-        rate.par = rate.par,

+        ratePar = ratePar,

         u = u,

         ermin = val,

         iter = iter

@@ -316,11 +316,11 @@ ClustComp <-

           NClust,

           m = m,

           verbose = FALSE,

-          iter.max =

+          iterMax =

             1000

         ))

     # cls <- lapply(seq_len(NSs), function(x) vsclust_algorithm(tData,NClust,

-    # m=m, verbose=FALSE,iter.max=1000))  #print(cls[[1]])

+    # m=m, verbose=FALSE,iterMax=1000))  #print(cls[[1]])

     Bestcl <- cls[[which.min(lapply(cls, function(x)

       x$withinerror))]]

     cls <-

@@ -330,7 +330,7 @@ ClustComp <-

           NClust,

           m = mm,

           verbose = FALSE,

-          iter.max =

+          iterMax =

             1000

         ))

     Bestcl2 <- cls[[which.min(lapply(cls, function(x)

@@ -39,7 +39,6 @@ print.fclust <-

 #' @param NClust Number of clusters

 #' @importFrom matrixStats rowMaxs

 #' @return fclust object with reorder clusters

-#' @example 

 #' @examples

 #' # Generate some random data

 #' data <- matrix(rnorm(seq_len(1000)), nrow=100)

@@ -83,9 +82,8 @@ SwitchOrder <- function(Bestcl, NClust) {

 #' @return Biological Homogeneity Index

 #' @examples

 #' # Run enrichment analysis

-#' library(clusterProfiler)

-#' data(gcSample)

-#' xx <- compareCluster(gcSample, fun="enrichKEGG",

+#' data(gcSample, package="clusterProfiler")

+#' xx <- clusterProfiler::compareCluster(gcSample, fun="enrichKEGG",

 #'                      organism="hsa", pvalueCutoff=0.05)

 #' # Generate random list from gcSample

 #' rand_ids <- lapply(gcSample, function(x) sample(unlist(gcSample), 200))

@@ -133,7 +131,7 @@ calcBHI <- function(Accs, gos) {

       ltgenes <- length(tgenes)

       if (ltgenes > 1) {

         for (i1 in tgenes[seq_len(ltgenes - 1)]) {

-          ttgene <- tgenes[(which(i1 == tgenes) + 1):ltgenes]

+          ttgene <- tgenes[seq(which(i1 == tgenes) + 1,ltgenes,1)]

           ispair[i1, ttgene] <- ispair[ttgene, i1] <- TRUE

         }

@@ -165,7 +163,7 @@ averageCond <- function(data, NumReps, NumCond) {

   # Calculates means over replicates

   tdat <-

     rowMeans(data[, seq(1, NumReps * NumCond, NumCond)], na.rm = TRUE)

-  for (i in 2:NumCond) {

+  for (i in seq(2,NumCond,1)) {

     tdat <-

       cbind(tdat, rowMeans(data[, seq(i, NumReps * NumCond, NumCond)], na.rm =

                              TRUE))

@@ -14,7 +14,7 @@ NULL

 #' scaling and filtering of missing values

 #' @param dat matrix or data frame of numerical data. Columns are samples.

 #' Replicates are grouped (i.e. A1, B1, C1, A2, B2, C2) when letters denote

-#' conditions and numbers the replicates. In case of `isStat=F`, you need a

+#' conditions and numbers the replicates. In case of `isStat=FALSE`, you need a

 #' last column for the standard deviations

 #' @param NumReps Number replicates in the data

 #' @param NumCond Number of different experimental conditions. The total number

@@ -123,7 +123,7 @@ PrepareForVSClust <-

 #' stats <- PrepareSEForVSClust(miniACC, coldatname="COC", isStat=TRUE)

 #'

 #' @import stats

-#' @import MultiAssayExperiment

+#' @importFrom MultiAssayExperiment assay assays sampleMap colData

 #' @importFrom matrixStats rowSds

 #' @importFrom shiny validate

 #' @export

@@ -287,7 +287,7 @@ estimClustNum <- function(dat,

   sds <- sds / (rowSds(as.matrix(tData), na.rm = TRUE))

   tData <- t(scale(t(tData)))

-  multiOut <- lapply(3:maxClust, function(x) {

+  multiOut <- lapply(seq(3,maxClust,1), function(x) {

     if (!is.null(getDefaultReactiveDomain())) {

       incProgress(1, detail = paste("Running cluster number", x))

     } else {

@@ -306,9 +306,9 @@ estimClustNum <- function(dat,

   stopCluster(cl)

-  for (NClust in 3:maxClust)

+  for (NClust in seq(3,maxClust,1))

     ClustInd[NClust - 2,] <- multiOut[[NClust - 2]]

-  rownames(ClustInd) <- paste0("num_clust_", 3:maxClust)

+  rownames(ClustInd) <- paste0("num_clust_", seq(3,maxClust,1))

   colnames(ClustInd) <-

     c(

       "MinCentroidDist_VSClust",

@@ -412,7 +412,7 @@ runClustWrapper <-

       tData,

       cl = Bestcl,

       mfrow = c(round(sqrt(NClust)), ceiling(sqrt(NClust))),

-      min.mem = 0.5,

+      minMem = 0.5,

       colo = "fancy"

     )

     p <- recordPlot()

@@ -9,7 +9,7 @@ PrepareForVSClust(dat, NumReps, NumCond, isPaired = FALSE, isStat)

 \arguments{

 \item{dat}{matrix or data frame of numerical data. Columns are samples.

 Replicates are grouped (i.e. A1, B1, C1, A2, B2, C2) when letters denote

-conditions and numbers the replicates. In case of `isStat=F`, you need a

+conditions and numbers the replicates. In case of `isStat=FALSE`, you need a

 last column for the standard deviations}

 \item{NumReps}{Number replicates in the data}

@@ -24,9 +24,8 @@ functional classes. BMC bioinformatics 7, 397 (2006).

 }

 \examples{

 # Run enrichment analysis

-library(clusterProfiler)

-data(gcSample)

-xx <- compareCluster(gcSample, fun="enrichKEGG",

+data(gcSample, package="clusterProfiler")

+xx <- clusterProfiler::compareCluster(gcSample, fun="enrichKEGG",

                      organism="hsa", pvalueCutoff=0.05)

 # Generate random list from gcSample

 rand_ids <- lapply(gcSample, function(x) sample(unlist(gcSample), 200))

@@ -11,7 +11,7 @@ estimClust.plot(ClustInd)

 }

 \value{

 Multiple panels showing expression profiles of clustered features 

-passing the min.mem threshold

+passing the minMem threshold

 }

 \description{

 This function visualizes the output from estimClustNumber, and there 

@@ -9,8 +9,8 @@ mfuzz.plot(

   cl,

   mfrow = c(1, 1),

   colo,

-  min.mem = 0,

-  time.labels,

+  minMem = 0,

+  timeLabels,

   filename = NA,

   xlab = "Time",

   ylab = "Expression changes"

@@ -27,10 +27,10 @@ panels are distributed in the plot}

 \item{colo}{color map to be used (can be missing)}

-\item{min.mem}{filter for showing only features with a higher membership 

+\item{minMem}{filter for showing only features with a higher membership 

 values than this value}

-\item{time.labels}{alternative labels for different conditions}

+\item{timeLabels}{alternative labels for different conditions}

 \item{filename}{for writing into pdf. Will write on screen when using NA}

@@ -40,7 +40,7 @@ values than this value}

 }

 \value{

 Multiple panels showing expression profiles of clustered features 

-passing the min.mem threshold

+passing the minMem threshold

 }

 \description{

 This function visualizes the clustered quantitative profiles in multiple 

@@ -52,7 +52,7 @@ labels and color maps. The code is adopted from the MFuzz package.

 data <- matrix(rnorm(seq_len(5000)), nrow=500)

 # Run clustering

 clres <- vsclust_algorithm(data, centers=2, m=1.5)

-mfuzz.plot(data, clres,  mfrow=c(2,3), min.mem=0.0)

+mfuzz.plot(data, clres,  mfrow=c(2,3), minMem=0.0)

 }

 \references{

 Schwaemmle V, Jensen ON. VSClust: feature-based variance-sensitive clustering 

@@ -7,11 +7,11 @@

 vsclust_algorithm(

   x,

   centers,

-  iter.max = 100,

+  iterMax = 100,

   verbose = FALSE,

   dist = "euclidean",

   m = 2,

-  rate.par = NULL,

+  ratePar = NULL,

   weights = 1,

   control = list()

 )

@@ -22,7 +22,7 @@ vsclust_algorithm(

 \item{centers}{Either numeric for number of clusters or numeric matrix with 

 center coordinates}

-\item{iter.max}{Numeric for maximum number of iterations}

+\item{iterMax}{Numeric for maximum number of iterations}

 \item{verbose}{Verbose information}

@@ -32,7 +32,7 @@ center coordinates}

 \item{m}{Fuzzifier value: numeric or vector of length equal to number of rows 

 of x}

-\item{rate.par}{(experimental) numeric value for punishing missing values}

+\item{ratePar}{(experimental) numeric value for punishing missing values}

 \item{weights}{numeric or vector of length equal to number of rows of x}

new file mode 100644

Binary files /dev/null and b/src/RcppExports.o differ

new file mode 100755

Binary files /dev/null and b/src/vsclust.so differ

@@ -326,7 +326,10 @@ double c_plusplus_means(const NumericMatrix & feature_mat, NumericMatrix & cente

                         NumericVector & weight, NumericVector & fuzz, int dist_metric, int iter_max, double rel_tol,

                         int verbose, NumericMatrix & membership_mat, double ermin, IntegerVector & iter, double missing_value = NA_REAL, 

                         double weight_missing = 0) {

-  

+

+  // check for user interrupts

+  Rcpp::checkUserInterrupt();

+

   int nr_objects = feature_mat.nrow();

   int nr_centers = centers.nrow();

   // for symmetric matrix

@@ -383,4 +386,4 @@ double c_plusplus_means(const NumericMatrix & feature_mat, NumericMatrix & cente

   ermin = new_fitness;

   return ermin;

-}

\ No newline at end of file

+}

new file mode 100644

Binary files /dev/null and b/src/vsclust_algo.o differ