YosefLab · deto · Feb 14, 2020 · Feb 6, 2020 · Feb 6, 2020 · Feb 14, 2020
diff --git a/DESCRIPTION b/DESCRIPTION
@@ -1,6 +1,6 @@
 Package: VISION
 Title: Functional interpretation of single cell RNA-seq latent manifolds
-Version: 2.0.0
+Version: 2.1.0
 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")),

diff --git a/NEWS.md b/NEWS.md
@@ -1,3 +1,17 @@
+# VISION 2.1.0
+
+Added parameter `sig_gene_threshold` with **changed default behavior**
+
+* Before it was a guideline to filter lowly expressed genes before running VISION
+* Now, by default, genes expressed in fewer than 0.1% of cells will be filtered automatically
+
+Bug Fixes:
+
+* Better colors in output when more than 10 categories
+* Errors with certain output object accessors
+* Selections not saving when loading results server
+* Crashes when running on more than 1200 signatures
+
 # VISION 2.0.0
 
 Lots of changes for this version.

diff --git a/R/AnalysisFunctions.R b/R/AnalysisFunctions.R
@@ -211,9 +211,11 @@ computeProjectionGenes <- function(object,
 #' on creating Signature objects.
 #' @param min_signature_genes Signature that match less than this number of genes in the
 #' supplied expression matrix are removed.
+#' @param sig_gene_threshold Proportion of cells that a gene must be detected in (nonzero)
+#' to be used in signature score calculations.
 #' @return the VISION object, with the @sigData slot updated
 #' @export
-addSignatures <- function(object, signatures, min_signature_genes=5) {
+addSignatures <- function(object, signatures, min_signature_genes=5, sig_gene_threshold=.01) {
 
     if (is.list(signatures)) {
         sigs <- lapply(signatures, function(sig){
@@ -237,7 +239,7 @@ addSignatures <- function(object, signatures, min_signature_genes=5) {
             Signature objects")
     }
 
-    sigs <- processSignatures(sigs, rownames(object@exprData), min_signature_genes)
+    sigs <- processSignatures(sigs, object@exprData, min_signature_genes, sig_gene_threshold)
 
     object@sigData <- c(object@sigData, sigs)
 

diff --git a/R/NormalizationMethods.R b/R/NormalizationMethods.R
@@ -110,17 +110,20 @@ getNormalizedCopySparse <- function(data, func) {
     if (func == "znorm_rows" || func == "znorm_rows_then_columns") {
         rowOffsets <- rowMeans(data) * -1
         rowScaleFactors <- rowVarsSp(data) ** -0.5
+        rowScaleFactors[is.infinite(rowScaleFactors)] <- 1
     }
 
     if (func == "znorm_columns") {
         colOffsets <- colMeans(data) * -1
         colScaleFactors <- colVarsSp(data) ** -0.5
+        colScaleFactors[is.infinite(colScaleFactors)] <- 1
     }
 
     if (func == "znorm_rows_then_columns") {
         result <- .colNormHelper(data, rowOffsets, rowScaleFactors)
         colOffsets <- result$colOffsets
         colScaleFactors <- result$colScaleFactors
+        colScaleFactors[is.infinite(colScaleFactors)] <- 1
     }
 
     nd <- NormData(data, rowOffsets = rowOffsets, colOffsets = colOffsets,

diff --git a/R/methods-Signature.R b/R/methods-Signature.R
@@ -4,11 +4,31 @@
 #' Drops signatures with less than `minSignatureGenes` matching genes
 #'
 #' @param sigData list of signature objects
-#' @param expressionGenes list of gene identifiers in the expression matrix
+#' @param exprData gene expression matrix
 #' @param minSignatureGenes minimum number of genes a signature must match
 #' in the expression matrix in order to be retained
+#' @param sig_gene_threshold Proportion of cells that a gene must be detected in (nonzero)
+#' to be used in signature score calculations.
+#' @importFrom Matrix rowSums
 #' @return processedSigData list of signature objects
-processSignatures <- function(sigData, expressionGenes, minSignatureGenes){
+processSignatures <- function(sigData, exprData, minSignatureGenes, sig_gene_threshold){
+    expressionGenes <- rownames(exprData)
+
+
+    cell_threshold <- sig_gene_threshold * ncol(exprData)
+    gene_detects <- rowSums(exprData > 0)
+    valid_genes <- gene_detects >= cell_threshold
+
+    message(
+        sprintf(
+            "\nUsing %i/%i genes detected in %.2f%% of cells for signature analysis.",
+            sum(valid_genes), nrow(exprData), sig_gene_threshold*100)
+        )
+    message(
+        "See the `sig_gene_threshold` input to change this behavior.\n"
+        )
+
+    expressionGenes <- rownames(exprData)[valid_genes]
     out <- lapply(sigData, function(sig){
         validGenes <- names(sig@sigDict) %in% expressionGenes
         sig@sigDict <- sig@sigDict[validGenes]

diff --git a/R/methods-Vision.R b/R/methods-Vision.R
@@ -20,6 +20,8 @@
 #' genes to use when computing projections.
 #' @param min_signature_genes Signature that match less than this number of genes in the
 #' supplied expression matrix are removed.
+#' @param sig_gene_threshold Proportion of cells that a gene must be detected in (nonzero)
+#' to be used in signature score calculations.
 #' @param threshold Threshold to apply when using the 'threshold' or 'fano' projection genes filter.
 #' If greater than 1, this specifies the number of cells in which a gene must be detected
 #' for it to be used when computing PCA. If less than 1, this instead specifies the proportion of cells needed
@@ -75,7 +77,9 @@ setMethod("Vision", signature(data = "matrixORSparse"),
             function(data, signatures=list(),
                     proteinData=NULL,
                     unnormalizedData = NULL, meta=NULL,
-                    projection_genes=c("fano"), min_signature_genes=5,
+                    projection_genes=c("fano"),
+                    min_signature_genes=5,
+                    sig_gene_threshold=.001,
                     threshold=.05, perm_wPCA=FALSE,
                     projection_methods = c("tSNE30"),
                     sig_norm_method = c("znorm_columns", "none", "znorm_rows",
@@ -101,7 +105,16 @@ setMethod("Vision", signature(data = "matrixORSparse"),
             .Object@params$micropooling <- list()
 
             rownames(data) <- toupper(rownames(data))
-            data <- data[ !duplicated(rownames(data)), , drop = FALSE]
+            toRemove <- rownames(data)[duplicated(rownames(data))]
+            if (length(toRemove) > 0){
+                message(sprintf(
+                        "\nRemoving %i genes with duplicate IDs (ignoring case): %s\n",
+                        length(toRemove) + length(unique(toRemove)),
+                        paste(unique(toRemove), collapse = ", ")
+                        )
+                    )
+                data <- data[ !(rownames(data) %in% toRemove), , drop = FALSE]
+            }
             .Object@exprData <- data
 
             if (!is.null(unnormalizedData)){
@@ -198,7 +211,7 @@ setMethod("Vision", signature(data = "matrixORSparse"),
                     Signature objects")
             }
 
-            .Object@sigData <- processSignatures(.Object@sigData, rownames(.Object@exprData), min_signature_genes)
+            .Object@sigData <- processSignatures(.Object@sigData, .Object@exprData, min_signature_genes, sig_gene_threshold)
 
             if (!is.null(meta)) {
                 if(is.matrix(meta)){

diff --git a/docs/LICENSE-text.html b/docs/LICENSE-text.html
diff --git a/docs/articles/Signatures.html b/docs/articles/Signatures.html
diff --git a/docs/articles/VISION-vignette.html b/docs/articles/VISION-vignette.html
diff --git a/docs/articles/index.html b/docs/articles/index.html
diff --git a/docs/articles/micropooling.html b/docs/articles/micropooling.html
diff --git a/docs/articles/trajectories.html b/docs/articles/trajectories.html
diff --git a/docs/articles/web_only/10xData.html b/docs/articles/web_only/10xData.html
diff --git a/docs/articles/web_only/Seurat.html b/docs/articles/web_only/Seurat.html
diff --git a/docs/articles/web_only/SignatureAutocorrelation.html b/docs/articles/web_only/SignatureAutocorrelation.html
diff --git a/docs/authors.html b/docs/authors.html
diff --git a/docs/index.html b/docs/index.html
diff --git a/docs/news/index.html b/docs/news/index.html
diff --git a/docs/reference/Cluster.html b/docs/reference/Cluster.html
diff --git a/docs/reference/NormData.html b/docs/reference/NormData.html
diff --git a/docs/reference/ServerExpression.html b/docs/reference/ServerExpression.html
diff --git a/docs/reference/ServerSigProjMatrix.html b/docs/reference/ServerSigProjMatrix.html
diff --git a/docs/reference/Signature.html b/docs/reference/Signature.html
diff --git a/docs/reference/Trajectory.html b/docs/reference/Trajectory.html
diff --git a/docs/reference/TrajectoryProjection.html b/docs/reference/TrajectoryProjection.html
diff --git a/docs/reference/VISION-class.html b/docs/reference/VISION-class.html
diff --git a/docs/reference/VISION.html b/docs/reference/VISION.html
diff --git a/docs/reference/addLatentSpace.html b/docs/reference/addLatentSpace.html