added background subtraction

nextflow.config

@@ -19,9 +19,6 @@ params {
     // Illumination correction
     illumination               = null
 
-    // Background subtraction
-    backsub                    = false
-
     // MultiQC options
     multiqc_config             = null
     multiqc_title              = null

nextflow_schema.json

@@ -74,10 +74,6 @@
                 "cellpose_model": {
                     "type": "string",
                     "description": "optional model file for cellpose segmentation"
-                },
-                "backsub": {
-                    "type": "boolean",
-                    "description": "boolean to flag whether or not to apply background subtraction"
                 }
             }
         },

subworkflows/local/utils_nfcore_mcmicro_pipeline/main.nf

@@ -221,6 +221,29 @@ def validateInputMarkersheet( markersheet_data ) {
         error("Duplicate [channel, cycle] pairs: ${dups}")
     }
 
+    // validate backsub columns if present
+    def exposure_list = markersheet_data.findResults{ _1, _2, _3, _4, _5, _6, exposure, _8, _9 -> exposure != [] ? exposure : null }
+    def background_list = markersheet_data.findResults{ _1, _2, _3, _4, _5, _6, _7, background, _9 -> background != [] ? background: null }
+    def remove_list = markersheet_data.findResults{ _1, _2, _3, _4, _5, _6, _7, _8, remove -> remove != [] ? remove : null }
+
+    if (background_list.size() == 0 && (exposure_list.size() > 0 || remove_list.size() > 0)) {
+        error("No values in background column, but values in either exposure or remove columns.  Must have background column values to perform background subtraction.")
+    } else if (background_list.size() > 0) {
+        inter_list = marker_name_list.intersect(background_list)
+        if (inter_list.size() != background_list.size()) {
+            outliers_list = background_list - inter_list
+            error('background column values must exist in the marker_name column. The following background column values do not exist in the marker_name column: ' + outliers_list)
+        }
+
+        if (exposure_list.size() == 0) {
+            error('You must have at least one value in the exposure column to perform background subtraction')
+        }
+
+        if (remove_list.size() == 0) {
+            error ('You must have at least one value in the remove column to perform background subtraction')
+        }
+    }
+
     return markersheet_data
 }
 

tests/main.nf.test

@@ -528,7 +528,6 @@ nextflow_workflow {
             when {
                 params {
                     segmentation = "mesmer"
-                    backsub = true
                 }
                 workflow {
                     """

workflows/mcmicro.nf

@@ -73,21 +73,80 @@ workflow MCMICRO {
         | ASHLAR
     ch_versions = ch_versions.mix(ASHLAR.out.versions)
 
-    // // Run Background Correction
-    if (params.backsub) {
-        ch_backsub_markers = ch_markersheet
-            .map { ['channel_number,cycle_number,marker_name,exposure,background,remove',
-                it.collect{ channel_number, cycle_number, marker_name, _1, _2, _3, exposure, background, remove ->
-                    channel_number + "," + cycle_number + "," + marker_name + "," + exposure + "," + background + "," + remove}] }
-            .flatten()
-            .map { it.replace('[]', '') }
-            .collectFile(name: 'markers_backsub.csv', sort: false, newLine: true, storeDir: "${workDir}")
-        BACKSUB(ASHLAR.out.tif, [[id:"$ASHLAR.out.tif[0]['id']"], "${workDir}/markers_backsub.csv"])
-        ch_segmentation_input = BACKSUB.out.backsub_tif
-        ch_versions = ch_versions.mix(BACKSUB.out.versions)
-    } else {
-        ch_segmentation_input = ASHLAR.out.tif
-    }
+    // Run Background Correction
+    ASHLAR_OUT_BACKSUB = ch_markersheet
+        .map { it.collect{ _1, _2, _3, _4, _5, _6, exposure, background, remove -> [exposure, background, remove]} }
+        .flatten()
+        .filter { it != [] }
+        .toList()
+        .combine(ASHLAR.out.tif)
+        .map { it.size() == 2 ? [it[0] + [backsub: false], [it[1]]] : [it[-2] + [backsub: true], it[-1]] }
+
+    ch_backsub_markers = ch_markersheet
+        .map { ['channel_number,cycle_number,marker_name,exposure,background,remove',
+            it.collect{ channel_number, cycle_number, marker_name, _1, _2, _3, exposure, background, remove ->
+                channel_number + "," + cycle_number + "," + marker_name + "," + exposure + "," + background + "," + remove}] }
+        .flatten()
+        .map { it.replace('[]', '') }
+        .collectFile(name: 'markers_backsub.csv', sort: false, newLine: true)
+
+    BACKSUB_INPUT = ASHLAR_OUT_BACKSUB
+        .combine(ch_backsub_markers)
+        .dump(tag: 'BACKSUB IN')
+        .multiMap{ meta, image, marker ->
+            image: [meta, image]
+            markers: [meta, marker]
+        }
+
+    BACKSUB_INPUT_IMAGE = BACKSUB_INPUT.image
+        .branch{ meta, image ->
+            exists: meta.backsub
+                return [meta, image]
+        }
+
+    BACKSUB_INPUT_MARKERS = BACKSUB_INPUT.markers
+        .branch{ meta, markers -> 
+            exists: meta.backsub
+                return [meta, markers]
+        }
+    BACKSUB(BACKSUB_INPUT_IMAGE.exists, BACKSUB_INPUT_MARKERS.exists)
+
+
+    /*  seems like this should replace the blocks (BACKSUB_INPUT, BACKSUB_INPUT_IMAGE and BACKSUB_INPUT_MARKERS) above.
+        from https://github.com/nextflow-io/nextflow/issues/1208 it seems like this should work, 
+        but as described using null fails 
+    ASHLAR_OUT_BACKSUB
+        .combine(ch_backsub_markers)
+        .dump(tag: 'BACKSUB IN')
+        .multiMap{ meta, image, marker ->
+            // image: meta.backsub ? [meta, image] : Channel.empty()
+            // markers: meta.backsub ? [meta, marker] : Channel.empty()
+            image: meta.backsub ? [meta, image] : null
+            markers: meta.backsub ? [meta, marker] : null
+            // image: meta.backsub ? [meta, image] : []
+            // markers: meta.backsub ? [meta, marker] : []
+        }
+        | BACKSUB
+    */
+
+    /*  another idea to replace the blocks (BACKSUB_INPUT, BACKSUB_INPUT_IMAGE and BACKSUB_INPUT_MARKERS) above.
+        would like to have a single branch condition return 2 channel outputs,
+        but I don't see how to do it. Is something like this possible?
+    BACKSUB_INPUT_TEST = ASHLAR_OUT_BACKSUB
+        .combine(ch_backsub_markers)
+        .branch{ meta, image, markers -> 
+            exists: meta.backsub
+                return [[meta1, image], [meta, markers]]
+        }
+        BACKSUB(BACKSUB_INPUT_TEST.exists)
+    */
+
+    ch_segmentation_input = ASHLAR_OUT_BACKSUB
+        .concat(BACKSUB.out.backsub_tif)
+        .map { [it[-2], it[-1]] }
+    ch_versions = ch_versions.mix(BACKSUB.out.versions)
+
+
 
     // Run Segmentation
 
@@ -119,6 +178,7 @@ workflow MCMICRO {
             ['marker_name'] +
             it.collect{ _1, _2, marker_name, _4, _5, _6, _7, _8, _9 -> '"' + marker_name + '"' }
         }
+        .dump(tag: "MARKERS")
         .collectFile(name: 'markers.csv', sort: false, newLine: true)
 
     ASHLAR.out.tif