Dont automatically ScaleData on all features

bbimber · bbimber · commit 4d79e377d5be · 2021-02-05T10:53:43.000-08:00
diff --git a/singlecell/resources/chunks/AppendCiteSeq.R b/singlecell/resources/chunks/AppendCiteSeq.R
@@ -1,5 +1,6 @@
 for (datasetId in names(seuratObjects)) {
     seuratObj <- seuratObjects[[datasetId]]
+    seuratObjects[[datasetId]] <- NULL
 
     if (!(datasetId %in% names(featureData))) {
         stop(paste0('No CITE-seq information found for datasetId: ', datasetId))
@@ -23,6 +24,5 @@ for (datasetId in names(seuratObjects)) {
     newSeuratObjects[[datasetId]] <- seuratObj
 
     # Cleanup
-    seuratObjects[[datasetId]] <- NULL
     gc()
 }
diff --git a/singlecell/resources/chunks/CiteSeqDimRedux.R b/singlecell/resources/chunks/CiteSeqDimRedux.R
@@ -1,11 +1,13 @@
 for (datasetId in names(seuratObjects)) {
     seuratObj <- seuratObjects[[datasetId]]
+    seuratObjects[[datasetId]] <- NULL
 
+    #TODO: what if not used?
     seuratObj <- CellMembrane::CiteSeqDimRedux(seuratObj)
 
     newSeuratObjects[[datasetId]] <- seuratObj
 
     # Cleanup
-    seuratObjects[[datasetId]] <- NULL
+    rm(seuratObj)
     gc()
 }
diff --git a/singlecell/resources/chunks/CiteSeqWnn.R b/singlecell/resources/chunks/CiteSeqWnn.R
@@ -1,11 +1,12 @@
 for (datasetId in names(seuratObjects)) {
     seuratObj <- seuratObjects[[datasetId]]
+    seuratObjects[[datasetId]] <- NULL
 
     seuratObj <- CellMembrane::RunSeuratWnn(seuratObj)
 
     newSeuratObjects[[datasetId]] <- seuratObj
 
     # Cleanup
-    seuratObjects[[datasetId]] <- NULL
+    rm(seuratObj)
     gc()
 }
diff --git a/singlecell/resources/chunks/DoubletFinder.R b/singlecell/resources/chunks/DoubletFinder.R
@@ -1,11 +1,12 @@
 for (datasetId in names(seuratObjects)) {
     seuratObj <- seuratObjects[[datasetId]]
+    seuratObjects[[datasetId]] <- NULL
 
     seuratObj <- CellMembrane::FindDoublets(seuratObj, dropDoublets = dropDoublets)
 
     newSeuratObjects[[datasetId]] <- seuratObj
 
     # Cleanup
-    seuratObjects[[datasetId]] <- NULL
+    rm(seuratObj)
     gc()
 }
diff --git a/singlecell/resources/chunks/Downsample.R b/singlecell/resources/chunks/Downsample.R
@@ -1,11 +1,12 @@
 for (datasetId in names(seuratObjects)) {
     seuratObj <- seuratObjects[[datasetId]]
+    seuratObjects[[datasetId]] <- NULL
 
     CellMembrane::DownsampleSeurat(seuratObj, targetCells = targetCells, subsetFields = subsetFields, seed = seed)
 
     newSeuratObjects[[datasetId]] <- seuratObj
 
     # Cleanup
-    seuratObjects[[datasetId]] <- NULL
+    rm(seuratObj)
     gc()
 }
diff --git a/singlecell/resources/chunks/FilterRawCounts.R b/singlecell/resources/chunks/FilterRawCounts.R
@@ -1,11 +1,12 @@
 for (datasetId in names(seuratObjects)) {
     seuratObj <- seuratObjects[[datasetId]]
+    seuratObjects[[datasetId]] <- NULL
 
     seuratObj <- bindArgs(CellMembrane::FilterRawCounts, seuratObj)()
 
     newSeuratObjects[[datasetId]] <- seuratObj
 
     # Cleanup
-    seuratObjects[[datasetId]] <- NULL
+    rm(seuratObj)
     gc()
 }
diff --git a/singlecell/resources/chunks/FindClustersAndDimRedux.R b/singlecell/resources/chunks/FindClustersAndDimRedux.R
@@ -1,11 +1,12 @@
 for (datasetId in names(seuratObjects)) {
     seuratObj <- seuratObjects[[datasetId]]
+    seuratObjects[[datasetId]] <- NULL
 
     seuratObj <- CellMembrane::FindClustersAndDimRedux(seuratObj, minDimsToUse = minDimsToUse)
 
     newSeuratObjects[[datasetId]] <- seuratObj
 
     # Cleanup
-    seuratObjects[[datasetId]] <- NULL
+    rm(seuratObj)
     gc()
 }
diff --git a/singlecell/resources/chunks/FindMarkers.R b/singlecell/resources/chunks/FindMarkers.R
@@ -1,10 +1,11 @@
 for (datasetId in names(seuratObjects)) {
     seuratObj <- seuratObjects[[datasetId]]
+    seuratObjects[[datasetId]] <- NULL
 
     dt <- CellMembrane::Find_Markers(seuratObj, identFields = identFields, outFile = paste0(datasetId, '.markers.txt'))
     print(dt)
 
     # Cleanup
-    seuratObjects[[datasetId]] <- NULL
+    rm(seuratObj)
     gc()
 }
diff --git a/singlecell/resources/chunks/NormalizeAndScale.R b/singlecell/resources/chunks/NormalizeAndScale.R
@@ -1,5 +1,4 @@
 for (datasetId in names(seuratObjects)) {
-    # Preemptively cleanup:
     seuratObj <- seuratObjects[[datasetId]]
     seuratObjects[[datasetId]] <- NULL
     gc()
@@ -9,6 +8,6 @@ for (datasetId in names(seuratObjects)) {
     newSeuratObjects[[datasetId]] <- seuratObj
 
     # Cleanup
-    seuratObjects[[datasetId]] <- NULL
+    rm(seuratObj)
     gc()
 }
diff --git a/singlecell/resources/chunks/PrepareRawCounts.R b/singlecell/resources/chunks/PrepareRawCounts.R
@@ -1,12 +1,13 @@
 for (datasetId in names(seuratObjects)) {
     seuratObj <- seuratObjects[[datasetId]]
+    seuratObjects[[datasetId]] <- NULL
 
     datasetName <- datasetIdToName[[datasetId]]
     seuratObj <- CellMembrane::ReadAndFilter10xData(dataDir = seuratObjects[[datasetId]], datasetId = datasetId, datasetName = datasetName)
 
     newSeuratObjects[[datasetId]] <- seuratObj
 
     # Cleanup
-    seuratObjects[[datasetId]] <- NULL
+    rm(seuratObj)
     gc()
 }
diff --git a/singlecell/resources/chunks/RemoveCellCycle.R b/singlecell/resources/chunks/RemoveCellCycle.R
@@ -1,11 +1,12 @@
 for (datasetId in names(seuratObjects)) {
     seuratObj <- seuratObjects[[datasetId]]
+    seuratObjects[[datasetId]] <- NULL
 
     seuratObj <- bindArgs(CellMembrane::RemoveCellCycle, seuratObj)()
 
     newSeuratObjects[[datasetId]] <- seuratObj
 
     # Cleanup
-    seuratObjects[[datasetId]] <- NULL
+    rm(seuratObj)
     gc()
 }
diff --git a/singlecell/resources/chunks/RunCellHashing.R b/singlecell/resources/chunks/RunCellHashing.R
@@ -1,5 +1,6 @@
 for (datasetId in names(seuratObjects)) {
     seuratObj <- seuratObjects[[datasetId]]
+    seuratObjects[[datasetId]] <- NULL
 
     if (!(datasetId %in% names(featureData))) {
         stop(paste0('No hashing information found for datasetId: ', datasetId))
@@ -17,6 +18,6 @@ for (datasetId in names(seuratObjects)) {
     newSeuratObjects[[datasetId]] <- seuratObj
 
     # Cleanup
-    seuratObjects[[datasetId]] <- NULL
+    rm(seuratObj)
     gc()
 }
diff --git a/singlecell/resources/chunks/RunPCA.R b/singlecell/resources/chunks/RunPCA.R
@@ -1,11 +1,12 @@
 for (datasetId in names(seuratObjects)) {
     seuratObj <- seuratObjects[[datasetId]]
+    seuratObjects[[datasetId]] <- NULL
 
-    seuratObj <- CellMembrane::RunPcaSteps(seuratObj, variableGenesWhitelist = variableGenesWhitelist, variableGenesBlacklist = variableGenesBlacklist, npcs = npcs)
+    seuratObj <- CellMembrane::RunPcaSteps(seuratObj, npcs = npcs)
 
     newSeuratObjects[[datasetId]] <- seuratObj
 
     # Cleanup
-    seuratObjects[[datasetId]] <- NULL
+    rm(seuratObj)
     gc()
 }
diff --git a/singlecell/resources/chunks/RunSingleR.R b/singlecell/resources/chunks/RunSingleR.R
@@ -1,11 +1,12 @@
 for (datasetId in names(seuratObjects)) {
     seuratObj <- seuratObjects[[datasetId]]
+    seuratObjects[[datasetId]] <- NULL
 
     seuratObj <- CellMembrane::RunSingleR(seuratObj)
 
     newSeuratObjects[[datasetId]] <- seuratObj
 
     # Cleanup
-    seuratObjects[[datasetId]] <- NULL
+    rm(seuratObj)
     gc()
 }
diff --git a/singlecell/resources/chunks/SplitSeurat.R b/singlecell/resources/chunks/SplitSeurat.R
@@ -1,12 +1,13 @@
 for (datasetId in names(seuratObjects)) {
     seuratObj <- seuratObjects[[datasetId]]
+    seuratObjects[[datasetId]] <- NULL
 
     newList <- CellMembrane::SplitSeurat(seuratObj, splitField = splitField, minCellsToKeep = minCellsToKeep)
     for (name in names(newList)) {
         newSeuratObjects[[paste0(datasetId, '-', name)]] <- newList[[name]]
     }
 
     # Cleanup
-    seuratObjects[[datasetId]] <- NULL
+    rm(seuratObj)
     gc()
 }
diff --git a/singlecell/resources/chunks/SubsetSeurat.R b/singlecell/resources/chunks/SubsetSeurat.R
@@ -1,11 +1,12 @@
 for (datasetId in names(seuratObjects)) {
     seuratObj <- seuratObjects[[datasetId]]
+    seuratObjects[[datasetId]] <- NULL
 
     seuratObj <- CellMembrane::SubsetSeurat(seuratObj, expressionStrings = expressionStrings)
 
     newSeuratObjects[[datasetId]] <- seuratObj
 
     # Cleanup
-    seuratObjects[[datasetId]] <- NULL
+    rm(seuratObj)
     gc()
 }
diff --git a/singlecell/src/org/labkey/singlecell/pipeline/singlecell/NormalizeAndScale.java b/singlecell/src/org/labkey/singlecell/pipeline/singlecell/NormalizeAndScale.java
@@ -25,7 +25,18 @@ public Provider()
                         put("initialValues", "vst");
                         put("allowBlank", false);
                     }}, "vst"),
-                    SeuratToolParameter.create("blockSize", "Block Size", "This will be passed to block.size in Seurat::ScaleData, which determines the number of features processed as a time. Increasing might increase speed at a memory cost, and descreasing on large datasets might reduce memory at a cost of overall speed.", "ldk-integerfield", new JSONObject(){{
+                    SeuratToolParameter.create("variableGenesWhitelist", "Genes to Add to VariableFeatures", "These genes, entered comma-separated or one/line, will be adding to the default Seurat::VariableFeatures gene set when running PCA", "sequenceanalysis-trimmingtextarea", new JSONObject(){{
+                        put("height", 150);
+                        put("delimiter", ",");
+                    }}, null),
+                    SeuratToolParameter.create("variableGenesBlacklist", "Genes to Exclude From VariableFeatures", "These genes, entered comma-separated or one/line, will be excluded from the genes passed to RunPCA (which is otherwise determined by Seurat::VariableFeatures)", "sequenceanalysis-trimmingtextarea", new JSONObject(){{
+                        put("height", 150);
+                        put("delimiter", ",");
+                    }}, null),
+                    SeuratToolParameter.create("scaleVariableFeaturesOnly", "ScaleData On Variable Features Only", "If checked, ScaleData will only be performed on VariableFeatures, which should dramatically reduce time and memory", "checkbox", new JSONObject(){{
+                        put("checked", true);
+                    }}, true),
+                    SeuratToolParameter.create("blockSize", "Block Size", "This will be passed to block.size in Seurat::ScaleData, which determines the number of features processed as a time. Increasing might increase speed at a memory cost, and decreasing on large datasets might reduce memory at a cost of overall speed.", "ldk-integerfield", new JSONObject(){{
 
                     }}, null, "block.size", false)
             ), Arrays.asList("/sequenceanalysis/field/TrimmingTextArea.js"), null);
diff --git a/singlecell/src/org/labkey/singlecell/pipeline/singlecell/RemoveCellCycle.java b/singlecell/src/org/labkey/singlecell/pipeline/singlecell/RemoveCellCycle.java
@@ -20,6 +20,9 @@ public static class Provider extends AbstractPipelineStepProvider<SingleCellStep
         public Provider()
         {
             super("RemoveCellCycle", "Remove Cell Cycle", "OOSAP", "This will score cells by cell cycle phase and regress out cell cycle.", Arrays.asList(
+                    SeuratToolParameter.create("scaleVariableFeaturesOnly", "ScaleData On Variable Features Only", "If checked, ScaleData will only be performed on VariableFeatures, which should dramatically reduce time and memory", "checkbox", new JSONObject(){{
+                        put("checked", true);
+                    }}, true),
                     SeuratToolParameter.create("blockSize", "Block Size", "This will be passed to block.size in Seurat::ScaleData, which determines the number of features processed as a time. Increasing might increase speed at a memory cost, and descreasing on large datasets might reduce memory at a cost of overall speed.", "ldk-integerfield", new JSONObject(){{
 
                     }}, null, "block.size", false)
diff --git a/singlecell/src/org/labkey/singlecell/pipeline/singlecell/RunPCA.java b/singlecell/src/org/labkey/singlecell/pipeline/singlecell/RunPCA.java
@@ -20,18 +20,10 @@ public static class Provider extends AbstractPipelineStepProvider<SingleCellStep
         public Provider()
         {
             super("RunPCA", "Run PCA", "OOSAP", "This will run standard Seurat PCA-related steps.", Arrays.asList(
-                    SeuratToolParameter.create("variableGenesWhitelist", "Genes to Add to VariableFeatures", "These genes, entered comma-separated or one/line, will be adding to the default Seurat::VariableFeatures gene set when running PCA", "sequenceanalysis-trimmingtextarea", new JSONObject(){{
-                        put("height", 150);
-                        put("delimiter", ",");
-                    }}, null),
-                    SeuratToolParameter.create("variableGenesBlacklist", "Genes to Exclude From VariableFeatures", "These genes, entered comma-separated or one/line, will be excluded from the genes passed to RunPCA (which is otherwise determined by Seurat::VariableFeatures)", "sequenceanalysis-trimmingtextarea", new JSONObject(){{
-                        put("height", 150);
-                        put("delimiter", ",");
-                    }}, null),
                     SeuratToolParameter.create("npcs", "NPCs", "Total Number of PCs to compute and store", "ldk-integerfield", new JSONObject(){{
                         put("minValue", 0);
                     }}, 50)
-            ), Arrays.asList("/sequenceanalysis/field/TrimmingTextArea.js"), null);
+            ), null, null);
         }
 
 
