Add support for additionalFieldsToAggregate

Author: bbimber

singlecell/resources/chunks/AvgExpression.R

@@ -15,7 +15,24 @@ GenerateAveragedData <- function(seuratObj, groupFields, addMetadata) {
         stop('When addMetadata=TRUE, cDNA_ID must be part of groupFields')
     }
 
-    a <- CellMembrane::PseudobulkSeurat(seuratObj, groupFields = groupFields, assays = assayName)
+    if (!all(is.null(additionalFieldsToAggregate))) {
+        if (any(grepl(additionalFieldsToAggregate, pattern = '\\*'))) {
+            additionalFieldsToAggregateExpanded <- c()
+            for (fn in additionalFieldsToAggregate) {
+                if (grepl(fn, pattern = '\\*')) {
+                    matchingCols <- grep(names(seuratObj@meta.data), pattern = fn, value = T)
+                    print(paste0('Expanding ', fn, ' into: ', paste0(matchingCols, collapse = ', ')))
+                    additionalFieldsToAggregateExpanded <- c(additionalFieldsToAggregateExpanded, matchingCols)
+                } else {
+                    additionalFieldsToAggregateExpanded <- c(additionalFieldsToAggregateExpanded, fn)
+                }
+
+                additionalFieldsToAggregate <- unique(additionalFieldsToAggregateExpanded)
+            }
+        }
+    }
+
+    a <- CellMembrane::PseudobulkSeurat(seuratObj, groupFields = groupFields, assays = assayName, additionalFieldsToAggregate = additionalFieldsToAggregate)
 
     if (addMetadata) {
         a <- Rdiscvr::QueryAndApplyMetadataUsingCDNA(a)

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

@@ -34,7 +34,13 @@ public Provider()
                     }}, true),
                     SeuratToolParameter.create("assayName", "Assay Name", "The assay to use", "textfield", new JSONObject(){{
 
-                    }}, "RNA")
+                    }}, "RNA"),
+                    SeuratToolParameter.create("additionalFieldsToAggregate", "Additional Field(s) to Aggregate", "Each field in this list must match a meta.data field. That field will be aggregated per group, and the mean reported in meta.data of the pseudobulk output object. Note: this also supports wildcards (which must contain asterisk). Any term containing asterisk will be interpreted as a regex and passed to grep() to find matching meta.data fields.", "sequenceanalysis-trimmingtextarea", new JSONObject(){{
+                        put("allowBlank", true);
+                        put("height", 150);
+                        put("delimiter", ",");
+                        put("stripCharsRe", "/['\"]/g");
+                    }}, "*UCell$").delimiter(",")
             ), null, null);
         }