Merge pull request #89 from YosefLab/yr-cass

PhyloVision & version bump

Author: mattjones315

DESCRIPTION

@@ -1,15 +1,17 @@
 Package: VISION
 Title: Functional interpretation of single cell RNA-seq latent manifolds
-Version: 2.1.0
+Version: 3.0.0
 Authors@R: c(person("Matt", "Jones", email = "mattjones315@gmail.com", role = c("aut", "cre")),
 			 person("David", "Detomaso", email = "david.detomaso@berkeley.edu", role = c("aut", "cre")),
 			 person("Tal", "Ashuach", email = "tal_ashuach@berkeley.edu", role = c("aut")),
+			 person("Yanay", "Rosen", email = "yanayrosen@berkeley.edu", role = c("aut")),
 			 person("Nir", "Yosef", email = "niryosef@berkeley.edu", role = c("ctb", "cph")))
 Author: Matt Jones [aut, cre], David Detomaso [aut, cre], Tal Ashuach [aut], Nir Yosef [ctb]
 Maintainer: Matt Jones <mattjones315@gmail.com>
 Description: VISION provides functional interpretation of single cell RNA-seq (scRNA-seq) latent manifolds through the use of biological signatures (which can be downloaded from online databases). VISION can operate downstream of other common analyses such as dimensionality reduction, clustering, or trajectory analysis of scRNA-seq data. VISION produces an interactive web-based output report that can be easily shared with other collaborators or the greater scientific community.
 Depends: R (>= 3.4)
 Imports:
+	dplyr,
 	fastICA,
 	igraph,
     irlba,
@@ -31,14 +33,16 @@ Imports:
 	Rtsne,
 	utils,
 	vegan,
+	ape,
+	reticulate,
 	wordspace
 LinkingTo: Rcpp
 License: MIT + file LICENSE
 Encoding: UTF-8
 LazyData: true
 URL: https://yoseflab.github.io/VISION, https://github.com/yoseflab/VISION
 BugReports: https://github.com/YosefLab/VISION/issues
-RoxygenNote: 6.1.1
+RoxygenNote: 7.1.1
 Suggests:
 	Biobase,
 	BiocStyle,

NAMESPACE

@@ -1,24 +1,49 @@
 # Generated by roxygen2: do not edit by hand
 
+export(PhyloVision)
 export(Vision)
+export(addHotspotToVision)
 export(addSignatures)
 export(addTSNE)
 export(addUMAP)
+export(analyzeHotspotObjectVision)
 export(analyzeLocalCorrelations)
+export(analyzeLocalCorrelationsModules)
 export(annotateLatentComponents)
 export(applyMicroClustering)
+export(calcModuleScores)
 export(calcSignatureScores)
+export(calc_mod_sig_enrichment)
+export(calc_set_enrichment)
+export(clusterModScores)
 export(clusterSigScores)
 export(computeLatentSpace)
 export(convertGeneIds)
 export(createGeneSignature)
+export(depthBasedCladewiseTreeCluster)
+export(depthBasedTreeCluster)
+export(draw_hotspot_heatmap)
+export(generateOverlapSignatures)
+export(group_modules_enrichment)
+export(hsCalculateModuleScores)
+export(hsComputeAutoCorrelations)
+export(hsComputeLocalCorrelations)
+export(hsCreateKnnGraph)
+export(hsInit)
+export(lca_based_depth)
+export(loadHotspotObject)
 export(poolMatrixCols)
 export(poolMatrixRows)
 export(poolMetaData)
 export(read_10x)
 export(read_10x_h5)
 export(read_10x_h5_v2)
 export(read_10x_h5_v3)
+export(runHotspot)
+export(saveHSBytestToPickle)
+export(treeClusterMinCladeSize)
+export(trivial_dist)
+exportMethods(PhyloVision)
 exportMethods(Vision)
 exportMethods(addProjection)
 exportMethods(analyze)
@@ -31,13 +56,16 @@ exportMethods(getSelections)
 exportMethods(getSignatureAutocorrelation)
 exportMethods(getSignatureDifferential)
 exportMethods(getSignatureScores)
+exportMethods(phyloAnalyze)
 exportMethods(saveAndViewResults)
 exportMethods(viewResults)
 import(Matrix)
 import(Rcpp)
+import(ape)
 import(loe)
 import(logging)
 import(methods)
+import(reticulate)
 importFrom(Matrix,Diagonal)
 importFrom(Matrix,Matrix)
 importFrom(Matrix,colMeans)

NEWS.md

@@ -1,3 +1,14 @@
+# VISION 3.0.0
+
+Major version bump!
+
+* Added support for Phylogenies as latent spaces in core VISION and integrated with VISION api
+* Re-engineered UI into Signature Autocorrelation and Hotspot views
+* Integrated [Hotspot](https://yoseflab.github.io/Hotspot/index.html) into VISION analysis and report UI.
+* Deprecated support for trajectories and LC Annotator.
+* New color-scheme for gene expression values
+
+
 # VISION 2.1.0
 
 Added parameter `sig_gene_threshold` with **changed default behavior**

R/AllClasses.R

@@ -5,10 +5,16 @@
 # scores, distance matrices, and anything else, is computed
 # on the different types of data.
 
+
 setClassUnion("numericORNULL", members = c("numeric", "NULL"))
 setClassUnion("matrixORSparse", members = c("matrix", "dgCMatrix"))
 setClassUnion("matrixORNULL", members = c("matrix", "NULL"))
 setClassUnion("dataframeORNULL", members = c("data.frame", "NULL"))
+setClassUnion("rawORNULL", members = c("raw", "NULL"))
+
+
+# setClassUnion("treeorNull", members=c("phylo", "NULL"))
+# setClassUnion("pythonorNull", members = c("python.builtin.object", "NULL"))
 
 Cluster <- setClass("Cluster",
     slots = c(
@@ -58,6 +64,8 @@ Trajectory <- setClass("Trajectory",
             # columns:  from (character), to (character), position (numeric, 0 to 1)
 ))
 
+setClassUnion("trajectoryORNULL", members = c("Trajectory", "NULL"))
+
 TrajectoryProjection <- setClass("TrajectoryProjection",
     slots = c(
         name = "character", # Name of projection
@@ -101,17 +109,23 @@ Vision <- setClass("Vision",
         unnormalizedData = "matrixORSparse",
         sigData = "list",
         metaData = "data.frame",
+        modData = "list",
         SigScores = "matrix",
+        ModScores = "matrix",
         LocalAutocorrelation = "list",
         TrajectoryAutocorrelation = "list",
         ClusterComparisons = "list",
         LCAnnotatorData = "LCAnnotatorDataOrNULL",
         Projections = "list",
         TrajectoryProjections = "list", # list of TrajectoryProjection
         SigGeneImportance = "list",
+        ModGeneImportance = "list",
         Pools = "list",
         LatentSpace = "matrix",
-        LatentTrajectory = "Trajectory",
+        LatentTrajectory = "trajectoryORNULL",
+        Hotspot = "rawORNULL",
+        ModuleSignatureEnrichment = "list",
+        ModuleHotspotScores = "data.frame",
         Viewer = "list",
         params = "list",
         version = "numeric"
@@ -122,18 +136,33 @@ Vision <- setClass("Vision",
         unnormalizedData = matrix(NA, 1, 1),
         sigData = list(),
         metaData = data.frame(),
+        modData=list(),
         SigScores = matrix(NA, 1, 1),
+        ModScores = matrix(NA, 1, 1),
         LocalAutocorrelation = list(),
         TrajectoryAutocorrelation = list(),
         ClusterComparisons = list(),
         LCAnnotatorData = NULL,
         Projections = list(),
         TrajectoryProjections = list(),
         SigGeneImportance = list(),
+        ModGeneImportance = list(),
         Pools = list(),
         LatentSpace = matrix(NA, 1, 1),
         LatentTrajectory = NULL,
+        Hotspot = NULL,
+        ModuleSignatureEnrichment = list(),
+        ModuleHotspotScores = data.frame(),
         Viewer = list(),
         params = list(),
         version = 1.2
 ))
+
+# This is a hack to try and get R to play nice with the phylo class
+# Which the Ape package doesn't expose
+phylo <- setClass("phylo")
+
+PhyloVision <- setClass("PhyloVision", contains = "Vision", 
+  slots = c(tree = "phylo"),
+  prototype = list(tree=NULL)
+)

R/AllGenerics.R

@@ -4,10 +4,20 @@ setGeneric("Vision", function(data, ...) {
     standardGeneric("Vision")
 })
 
+#' @rdname PhyloVision-class
+#' @export
+setGeneric("PhyloVision", function(tree, ...) {
+  standardGeneric("PhyloVision")
+})
+
 setGeneric("analyze", function(object, ...) {
     standardGeneric("analyze")
 })
 
+setGeneric("phyloAnalyze", function(object, ...) {
+  standardGeneric("phyloAnalyze")
+})
+
 setGeneric("addProjection", function(object, ...) {
     standardGeneric("addProjection")
 })

R/AnalysisFunctions.R

@@ -4,21 +4,32 @@
 #'
 #' @param object the VISION object for which to cluster the cells
 #' @return the VISION object modifed as described above
-clusterCells <- function(object) {
+clusterCells <- function(object, tree=FALSE) {
 
     message("Clustering cells...", appendLF = FALSE)
 
     res <- object@LatentSpace
 
     K <- min(object@params$numNeighbors, 30)
-    kn <- find_knn_parallel(res, K)
-
-    cl <- louvainCluster(kn, res)
-
+    
+    if (!tree) {
+      kn <- find_knn_parallel(res, K)
+      cl <- louvainCluster(kn, res)
+    } else {
+        message("Using Tree to compute clusters...\n")
+        # Get the MRCA matrix and convert the node indexes to depths
+        cl <- maxSizeCladewiseTreeCluster(object@tree)
+    }
+    
     names(cl) <- paste('Cluster', seq(length(cl)))
 
     # cl is list of character vector
-    cluster_variable <- "VISION_Clusters"
+    if (!tree) {
+      cluster_variable <- "VISION_Clusters"
+    } else {
+      cluster_variable <- "VISION_Clusters_Tree"
+    }
+    
     metaData <- object@metaData
 
     metaData[cluster_variable] <- factor(levels = names(cl))
@@ -154,7 +165,7 @@ computeProjectionGenes <- function(object,
     if (!is.null(projection_genes)){
         object@params$latentSpace$projectionGenes <- projection_genes
     } else {
-        object@params$latentSpace$projectionGenes <- NULL
+        object@params$latentSpace$projectionGenes <- NA
     }
 
     if (!is.null(projection_genes_method)){
@@ -163,7 +174,7 @@ computeProjectionGenes <- function(object,
 
     message("Determining projection genes...")
 
-    if (is.null(object@params$latentSpace[["projectionGenes"]])){
+    if (is.na(object@params$latentSpace[["projectionGenes"]])){
 
         exprData <- matLog2(object@exprData)
         projection_genes <- applyFilters(
@@ -283,22 +294,43 @@ calcSignatureScores <- function(
         object@exprData,
         object@params$signatures$sigNormMethod
     )
-
-    sigScores <- batchSigEvalNorm(object@sigData, normExpr)
-
+    
+    sigData <- object@sigData
+    
+    for (sig_name in names(object@sigData)){
+        signature <- object@sigData[[sig_name]]
+        directional <- all(c(1, -1) %in% signature@sigDict)
+        if (directional) {
+            up <- names(which(signature@sigDict == 1))
+            down <- names(which(signature@sigDict == -1))
+            
+            up_name <- paste(signature@name, "_UP", sep = "")
+            down_name <- paste(signature@name, "_DOWN", sep = "")
+            
+            up_sig <- Signature(sigDict = signature@sigDict[up], name = up_name, source = signature@source, metaData = signature@metaData)
+            down_sig <- Signature(sigDict = signature@sigDict[down], name = down_name, source = signature@source, metaData = signature@metaData)
+            sigData[[up_name]] <- up_sig
+            sigData[[down_name]] <- down_sig
+        }
+    }
+    
+    object@sigData <- sigData # persist back to the object because of issues with up/down
+    
+    sigScores <- batchSigEvalNorm(sigData, normExpr)
+    
     if (sig_gene_importance) {
 
         if (is(object@exprData, "sparseMatrix")) {
             sigGeneImportance <- evalSigGeneImportanceSparse(
-                sigScores, object@sigData, normExpr
+                sigScores, sigData, normExpr
             )
         } else {
             normExprDense <- getNormalizedCopy(
                 object@exprData,
                 object@params$signatures$sigNormMethod
             )
             sigGeneImportance <- evalSigGeneImportance(
-                sigScores, object@sigData, normExprDense
+                sigScores, sigData, normExprDense
             )
         }
 
@@ -715,7 +747,7 @@ addTSNE <- function(object, perplexity = 30, name = "tSNE", source = "LatentSpac
 #' @param object the VISION object
 #' @return the VISION object with values set for the analysis results
 #' @export
-analyzeLocalCorrelations <- function(object) {
+analyzeLocalCorrelations <- function(object, tree=FALSE) {
 
   signatureBackground <- generatePermutationNull(
       object@exprData, object@sigData, num = 3000
@@ -726,9 +758,13 @@ analyzeLocalCorrelations <- function(object) {
       object@params$signatures$sigNormMethod)
 
   message("Computing KNN Cell Graph in the Latent Space...\n")
-
-  weights <- computeKNNWeights(object@LatentSpace, object@params$numNeighbors)
-
+  if (!tree) {
+    weights <- computeKNNWeights(object@LatentSpace, object@params$numNeighbors)
+  } else {
+    message("Using Tree to compute neighbors...\n")
+    weights <- computeKNNWeights(object@tree, object@params$numNeighbors)
+  }
+ 
   message("Evaluating local consistency of signatures in latent space...\n")
 
   sigConsistencyScores <- sigConsistencyScores(