Switch from Seurat AvgExpression to Pseudobulk

Author: bbimber

singlecell/resources/chunks/AvgExpression.R

@@ -15,25 +15,7 @@ GenerateAveragedData <- function(seuratObj, groupFields, addMetadata) {
         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 = '_'))
-    })
-
-    Seurat::Idents(seuratObj) <- rownames(meta)
-
-    for (assayName in names(seuratObj@assays)) {
-        if (!(!identical(seuratObj@assays[[assayName]]@counts, seuratObj@assays[[assayName]]@data))){
-            print(paste0('Seurat assay', assayName, ' does not appear to be normalized, running now:'))
-            seuratObj <- Seurat::NormalizeData(seuratObj, verbose = FALSE, assay = assayName)
-        }
-    }
-
-    a <- Seurat::AverageExpression(seuratObj, return.seurat = T, verbose = F)
-    a <- Seurat::AddMetaData(a, meta)
-
-    totals <- seuratObj@meta.data %>% group_by_at(groupFields) %>% summarise(TotalCells = n())
-    a$TotalCells <- totals$TotalCells
+    a <- CellMembrane::PseudobulkSeurat(seuratObj, groupFields = groupFields, assays = assayName)
 
     if (addMetadata) {
         a <- Rdiscvr::QueryAndApplyMetadataUsingCDNA(a)

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

@@ -22,7 +22,7 @@ public static class Provider extends AbstractPipelineStepProvider<SingleCellStep
     {
         public Provider()
         {
-            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(
+            super("AvgExpression", "Pseudobulk", "CellMembrane/Seurat", "This will run Pseudobulking on the raw counts, grouping using the provided fields. It will generate a seurat object with sum of counts per group/feature.", 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.", "sequenceanalysis-trimmingtextarea", new JSONObject(){{
                         put("allowBlank", false);
                         put("height", 150);
@@ -31,7 +31,10 @@ public Provider()
                     }}, "cDNA_ID").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)
+                    }}, true),
+                    SeuratToolParameter.create("assayName", "Assay Name", "The assay to use", "textfield", new JSONObject(){{
+
+                    }}, "RNA")
             ), null, null);
         }