Add simple function to run AverageSeurat

Author: bbimber

singlecell/resources/chunks/AvgExpression.R

@@ -1,10 +1,35 @@
+GenerateAveragedData <- function(seuratObj, groupFields, addMetadata) {
+    if (addMetadata && !'cDNA_ID' %in% names(seuratObj@meta.data)) {
+        stop('A field names cDNA_ID must exist when addMetadata=TRUE')
+    }
+
+    if (addMetadata && !'cDNA_ID' %in% groupFields) {
+        stop('When addMetadata=TRUE, cDNA_ID must be part of groupFields')
+    }
+
+    meta <- unique(seuratObj@meta.data[,groupFields, drop = F])
+    rownames(meta) <- apply(meta, 1, function(y){
+        return(paste0(y, collapse = '_'))
+    })
+
+    Idents(seuratObj) <- rownames(meta)
+
+    a <- Seurat::AverageExpression(seuratObj, return.seurat = T, verbose = F)
+    a <- Seurat::AddMetaData(a, meta)
+
+    if (addMetadata) {
+        a <- Rdiscvr::QueryAndApplyMetadataUsingCDNA(a)
+    }
+
+    return(a)
+}
+
 for (datasetId in names(seuratObjects)) {
     printName(datasetId)
     seuratObj <- readRDS(seuratObjects[[datasetId]])
 
-    df <- CellMembrane::AvgExpression(seuratObj, groupField = groupField)
-    write.table(df, file = paste0(outputPrefix, '.', makeLegalFileName(datasetId), '.avg.', groupField, '.txt'), sep = '\t', row.names = FALSE, quote = FALSE)
-    rm(df)
+    seuratObj <- GenerateAveragedData(seuratObj, groupFields = groupFields, addMetadata = addMetadata)
+    saveData(seuratObj, datasetId)
 
     # Cleanup
     rm(seuratObj)

singlecell/src/org/labkey/singlecell/pipeline/singlecell/AvgExpression.java

@@ -8,7 +8,7 @@
 
 import java.util.Arrays;
 
-public class AvgExpression extends AbstractCellMembraneStep
+public class AvgExpression extends AbstractRDiscvrStep
 {
     public AvgExpression(PipelineContext ctx, AvgExpression.Provider provider)
     {
@@ -19,10 +19,15 @@ public static class Provider extends AbstractPipelineStepProvider<SingleCellStep
     {
         public Provider()
         {
-            super("AvgExpression", "Avg. Expression", "Seurat", "This will run AverageExpression on the raw counts, producing a matrix with the average per group. This matrix will have a row labeled TotalCells appended, which is the total cells per group.", Arrays.asList(
-                    SeuratToolParameter.create("groupField", "Grouping Field", "This field will be used to group cells of the seurat object. For each unique value of this field, count averages will be computed and saved into a matrix with one column per group. Any cells lacking a value in this field will be discarded.", "textfield", new JSONObject(){{
+            super("AvgExpression", "Avg. Expression", "Seurat", "This will run AverageExpression on the raw counts, grouping using the provided fields. It will generate a seurat object with aggregate counts.", Arrays.asList(
+                    SeuratToolParameter.create("groupFields", "Grouping Field(s)", "This field will be used to group cells of the seurat object. For each unique value of this field, count averages will be computed and saved into a matrix with one column per group. Any cells lacking a value in this field will be discarded.", "textfield", new JSONObject(){{
                         put("allowBlank", false);
-                    }}, null)
+                        put("height", 150);
+                        put("delimiter", ",");
+                    }}, null).delimiter(","),
+                    SeuratToolParameter.create("addMetadata", "Query Metadata?", "If checked, Rdiscvr::QueryAndApplyMetadataUsingCDNA will be run after aggregation. This requires a cDNA_ID column to exist.", "checkbox", new JSONObject(){{
+                        put("checked", true);
+                    }}, true)
             ), null, null);
         }
 
@@ -42,7 +47,7 @@ public String getFileSuffix()
     @Override
     public boolean createsSeuratObjects()
     {
-        return false;
+        return true;
     }
 }