Merge branch 'release-1.0.1' into master

.Rbuildignore

@@ -20,4 +20,5 @@ runtime.txt
 .travis.yml
 .github
 .git
-Aux
+Aux
+binder

.github/ISSUE_TEMPLATE/bug_report.md

@@ -1,38 +1,28 @@
 ---
 name: Bug report
-about: Create a report to help us improve
+about: Create a report to help us improve DIscBIO
 title: ''
 labels: bug
 assignees: ''
 
 ---
 
-# Describe the bug
+# Summary
 A clear and concise description of what the bug is.
 
-# Steps to Reproduce
+# DIscBIO version
+Type `package.version("DIscBIO")` and paste the output here.
 
-Try to create a minimally-reproducible example so other people can better understand the problem.
+# Expected output
+What did you expect to happen?
 
-Steps to reproduce the behavior:
-1. Go to '...'
-2. Click on '....'
-3. Scroll down to '....'
-4. Reproduce the following code:
-```
-
-```
+# Obtained output
+What happened instead?
 
-# Expected behavior
-A clear and concise description of what you expected to happen.
+# Steps to reproduce issue
 
-# Screenshots
-If applicable, add screenshots to help explain your problem.
+Paste a script that reproduces the error inside the "```" marks.
 
-# Software metainformation
- - Operating System and version: [e.g. Linux, Mac, Windows]
- - R version:
- - DIscBIO version or commit number:
+```r
 
-# Additional context
-Add any other context about the problem here.
+```

.travis.yml

@@ -4,6 +4,10 @@ cache:
   apt: true
   packages: true
   timeout: 3000
+before_install:
+  - sudo apt-get update
+  - sudo apt-get -y install openjdk-9-jre-headless openjdk-9-source
+  - sudo apt-get -y install r-cran-rjava
 r:
   - release
   - devel

DESCRIPTION

@@ -1,7 +1,7 @@
 Package: DIscBIO
-Date: 2020-07-30
+Date: 2020-10-06
 Title: A User-Friendly Pipeline for Biomarker Discovery in Single-Cell Transcriptomics
-Version: 1.0.0
+Version: 1.0.1
 Authors@R:
 	c(
 		person(
@@ -48,7 +48,7 @@ Suggests:
 	Seurat
 LazyData: true
 RoxygenNote: 7.1.1
-URL: https://github.com/ocbe-uio/DIscBIO#usage
+URL: https://github.com/ocbe-uio/DIscBIO
 BugReports: https://github.com/ocbe-uio/DIscBIO/issues
 Collate:
     'DIscBIO-classes.R'

CHANGELOG.md → NEWS.md

@@ -1,10 +1,16 @@
 This file documents the changes between stable releases of DIscBIO.
 
-The next stable release of DIscBIO will contain several changes from the [original work](https://github.com/SystemsBiologist/PSCAN), the most relevant of which are listed below:
+# Version 1.0.1
+
+- Bug fixes
+- Updated Binder notebook
+
+# Version 1.0.0
+
+Contains several changes from the [original work](https://github.com/SystemsBiologist/PSCAN), the most relevant of which are listed below:
 
 - Several functions have been enhanced with a `quiet` argument to suppress intermediate output (essential to obtain cleaner output in batch scripts and unit tests)
 - Functions have been isolated into their own R files
 - Function and package documentation were added
 - Test datasets were compressed to occupy less disk space
-- The accompanying [Jupyter notebook](https://hub.gke.mybinder.org/user/systemsbiologist-pscan-tza36hyq/notebooks/DIscBIO.ipynb) was integrated as a vignette
 - Bugs have been fixed

R/DIscBIO-classes.R

@@ -151,8 +151,6 @@ setMethod(
         }
 
         # Proceed
-        #tmpFeats <- rownames(expdataAll)
-        #tmpExpdataAll <- expdataAll
         tmpFeats <- rownames(.Object@expdataAll)
         tmpExpdataAll <- .Object@expdataAll
 

R/DIscBIO-generic-ClustDiffGenes.R

@@ -60,9 +60,6 @@ setMethod(
         } else if (pValue < 0 | pValue > 1) {
             stop("pValue has to be a number between 0 and 1")
         }
-        if (length(object@kmeans$kpart) == 0) {
-            stop("run Clustexp before ClustDiffGenes")
-        }
         cdiff <- list()
         x     <- object@ndata
         y     <- object@expdata[, names(object@ndata)]
@@ -71,7 +68,7 @@ setMethod(
         } else if (ran_m) {
             part  <- object@MBclusters$clusterid
         } else {
-            stop("Run Clustexp() before running this function")
+            stop("Run clustering before running this function")
         }
         # ======================================================================
         # Operating
@@ -248,4 +245,4 @@ setMethod(
             print(paste("There are no DEGs with fdr =", fdr))
         }
     }
-)
+)

R/DIscBIO-generic-FindOutliers.R

@@ -98,8 +98,9 @@ setMethod(
             outdistquant = outdistquant
         )
         ### calibrate background model
-        m <- log2(apply(object@fdata, 1, mean))
-        v <- log2(apply(object@fdata, 1, var))
+	EXP<-object@expdata + 0.1
+        m <- log2(apply(EXP, 1, mean))
+        v <- log2(apply(EXP, 1, var))
         f <- m > -Inf & v > -Inf
         m <- m[f]
         v <- v[f]
@@ -259,7 +260,7 @@ setMethod(
                         y,
                         ylim = c(0, max(-log10(
                             x + 2.2e-16
-                        )) * 1.1),
+                        )) * 2.1),
                         col = fcol[i],
                         border = fcol[i],
                         names.arg = FALSE,
@@ -299,4 +300,4 @@ setMethod(
         LL <- which(object@cpart > K)
         return(LL)
     }
-)
+)

R/DIscBIO-generic-clusteringOrder.R

@@ -7,6 +7,9 @@
 #' @param export if `TRUE`, exports order table to csv
 #' @param filename Name of the exported file (if `export=TRUE`)
 #' @importFrom TSCAN TSCANorder
+#' @note This function has been replaced by pseudoTimeOrdering(), but it is
+#'   being kept for legacy purposes. It will, however, be removed from future
+#'   versions of DIscBIO.
 #' @return The DISCBIO-class object input with the kordering slot filled.
 setGeneric("KmeanOrder", function(object, quiet = FALSE, export = FALSE,
     filename = "Cellular_pseudo-time_ordering_based_on_k-meansc-lusters")
@@ -19,6 +22,15 @@ setMethod(
     "KmeanOrder",
     signature = "DISCBIO",
     definition = function(object, quiet, export, filename) {
+        # TODO: update Notebook not to depend on pseudoTimeOrdering instead
+        warning(
+            "KmeanOrder() has been replaced with pseudoTimeOrdering(), ",
+            "which performs pseudo-time ordering for both k-means ",
+            "and model-based clustering. ",
+            "This function is being kept for legacy purposes, ",
+            "but will be removed in future versions of DIscBIO. ",
+            "Please adapt your scripts accordingly."
+        )
         # Validation
         if (length(object@kmeans$kpart) == 0) {
             stop("run Clustexp before KmeanOrder")

R/DIscBIO-generic-clustheatmap.R

@@ -57,7 +57,6 @@ setMethod(
 			object@distances <- as.matrix(y$di)
     		part <- object@MBclusters$clusterid
 		}
-        part <- object@kmeans$kpart
         na <- c()
         j <- 0
         for (i in 1:max(part)) {
@@ -172,4 +171,4 @@ setMethod(
         }
         return(cclmo)
     }
-)
+)

R/DIscBIO-generic-plotExptSNE.R

@@ -41,7 +41,8 @@ setMethod(
         # ======================================================================
         # Plotting
         # ======================================================================
-        l <- apply(object@ndata[g, ] - .1, 2, sum) + .1
+        logObj <- log(object@ndata)
+        l <- apply(logObj[g, ] - .1, 2, sum) + .1
         mi <- min(l, na.rm = TRUE)
         ma <- max(l, na.rm = TRUE)
         ColorRamp <- colorRampPalette(
@@ -61,7 +62,7 @@ setMethod(
         opar <- par(mar = c(3, 5, 2.5, 2))
         on.exit(par(opar))
         plot(
-            object@tsne,
+            x,
             xlab = "Dim 1",
             ylab = "Dim 2",
             main = n,
@@ -97,4 +98,4 @@ setMethod(
         )
         layout(1)
     }
-)
+)

R/DIscBIO-generic-plotSilhouette.R

@@ -26,11 +26,16 @@ setMethod(
         # Validation
         # ======================================================================
         ran_clustexp <- length(object@kmeans$kpart) > 0
-        ran_exprmclust <- length(object@MBclusters$clusterid) > 0
+        ran_exprmclust <- length(object@MBclusters) > 0
         if (ran_clustexp) {
             kpart <- object@kmeans$kpart
+            DIS <- object@distances
         } else if (ran_exprmclust) {
             kpart <- object@MBclusters$clusterid
+            y <- clustfun(object@fdata, clustnr = 3, bootnr = 50,
+                 metric = "pearson", do.gap = TRUE, SE.method = "Tibs2001SEmax",
+                 SE.factor = .25, B.gap = 50, cln = 0, rseed = NULL, quiet = TRUE )
+            DIS <- as.matrix(y$di)
         } else {
             stop("run clustexp or exprmclust before plotSilhouette")
         }
@@ -41,8 +46,8 @@ setMethod(
         # Plotting
         # ======================================================================
         col <- c("black", "blue", "green", "red", "yellow", "gray")
-        distances <- dist.gen(object@distances)
+        distances <- DIS
         si <- silhouette(kpart, distances)
         plot(si, col = col[1:K])
     }
-)
+)

R/DIscBIO-generic-plottSNE.R

@@ -30,6 +30,7 @@ setMethod(
         # Plotting
         # ======================================================================
         col <- c("black", "blue", "green", "red", "yellow", "gray")
+	LEN<-length(levels(factor(part)))
         plot(
             x,
             las = 1,
@@ -39,7 +40,7 @@ setMethod(
             cex = 1.5,
             col = "lightgrey"
         )
-        for (i in seq_len(part)) {
+        for (i in seq_len(LEN)) {
             if (sum(part == i) > 0) {
                 text(
                     x[part == i, 1],

R/DIscBIO-generic-pseudoTimeOrdering.R

@@ -55,7 +55,11 @@ setMethod(
         FinalOrder <- orderTable[match(sampleNames, orderTable$orderID), ]
         out_order <- FinalOrder[, 1]
         names(out_order) <- names(Names)
-        object@kordering <- out_order
+        if (ran_k) {
+            object@kordering <- out_order
+        } else if (ran_m) {
+            object@MBordering <- out_order
+        }
         return(object)
     }
-)
+)

R/J48DTeval.R

@@ -74,18 +74,18 @@ J48DTeval <- function(
         )
         colnames(j48.confusion.matrix) <- c(First, Second)
         if (!quiet) print(j48.confusion.matrix)
-        j48.sn <- SN(j48.confusion.matrix)
-        j48.sp <- SP(j48.confusion.matrix)
-        j48.acc <- ACC(j48.confusion.matrix)
-        j48.mcc <- MCC(j48.confusion.matrix)
+        j48.sn <- round(SN(j48.confusion.matrix),digits=2)
+        j48.sp <- round(SP(j48.confusion.matrix),digits=2)
+        j48.acc <- round(ACC(j48.confusion.matrix),digits=2)
+        j48.mcc <- round(MCC(j48.confusion.matrix),digits=2)
 
         if (!quiet) {
             message(
                 "J48 SN: ", j48.sn, "\n",
                 "J48 SP: ", j48.sp, "\n",
                 "J48 ACC: ", j48.acc, "\n",
-                "J48 MCC: ", j48.mcc, "\n",
+                "J48 MCC: ", j48.mcc, "\n"
             )
         }
         return(j48.performance)
-    }
+    }

R/Jaccard.R

@@ -26,9 +26,6 @@ Jaccard <- function(
     JACCARD <- c()
 
     # Validation
-    if (length(object@kmeans$kpart) == 0) {
-        stop("run Clustexp before Jaccard")
-    }
     if (!(Clustering %in% c("K-means", "MB"))) {
         stop("Clustering has to be either K-means or MB")
     }
@@ -71,4 +68,4 @@ Jaccard <- function(
         box()
     }
     return(JACCARD)
-}
+}

R/Networking.R

@@ -50,7 +50,7 @@ Networking <- function(
                 "Examine response components =",
                 status_code(repos),
                 "\t",
-                "(200 means successful)",
+                "(200 means successful)"
             )
             y <- repos$request$url
             if (!is.null(FileName)) {
@@ -74,11 +74,11 @@ Networking <- function(
             rasterImage(Network, 0, 0, 1, 1)
             message(
                 "\n",
-                "You can see the network with high resolution",
+                "You can see the network with high resolution ",
                 "by clicking on the following link:",
                 "\n",
                 paste0(y)
             )
             set_plot_dimensions(8, 8) # resets to default values
         }
-    }
+    }

R/PlotMBpca.R

@@ -32,7 +32,8 @@ PlotMBpca <- function(object, type="order", g=NULL, n=NULL) {
         if (is.null(n)) {
             n <- g[1]
         }
-        l <- apply(object@ndata[g,] - .1, 2, sum) + .1
+        logObj <- log(object@ndata)
+        l <- apply(logObj[g,] - .1, 2, sum) + .1
         x <- data$pcareduceres
     } else if (type == "order") {
         MBordertable <- cbind(data$pcareduceres, object@MBordering)
@@ -97,4 +98,4 @@ PlotMBpca <- function(object, type="order", g=NULL, n=NULL) {
         xaxt = "n"
     )
     layout(1)
-}
+}