@platforma-open/milaboratories.top-antibodies.workflow 4.0.0 → 4.0.2

package/.turbo/turbo-build.log

@@ -1,6 +1,6 @@
  WARN  Issue while reading "/home/runner/work/antibody-tcr-lead-selection/antibody-tcr-lead-selection/.npmrc". Failed to replace env in config: ${NPMJS_TOKEN}
 
-> @platforma-open/milaboratories.top-antibodies.workflow@4.0.0 build /home/runner/work/antibody-tcr-lead-selection/antibody-tcr-lead-selection/workflow
+> @platforma-open/milaboratories.top-antibodies.workflow@4.0.2 build /home/runner/work/antibody-tcr-lead-selection/antibody-tcr-lead-selection/workflow
 > shx rm -rf dist && pl-tengo check && pl-tengo build
 
 Processing "src/assembling-fasta.tpl.tengo"...

package/CHANGELOG.md

@@ -1,5 +1,23 @@
 # @platforma-open/milaboratories.top-antibodies.workflow
 
+## 4.0.2
+
+### Patch Changes
+
+- 2a2533d: Fix minor issues
+- 6042e4a: Minor fix
+- 461999c: Fix minor issues
+
+## 4.0.1
+
+### Patch Changes
+
+- dd754ae: Accept both pre- and post-peptide-adaptation spec names from upstream blocks so projects using either version remain functional:
+
+  - Preset filter/ranking allowlists now include `pl7.app/enrichment*` (clonotype-enrichment) and `pl7.app/developability*` (antibody-sequence-liabilities) alongside the legacy `pl7.app/vdj/`-prefixed names.
+  - Diversification dropdown, cluster-axis matching, hidden cluster-mapping column, and workflow-side linker matching now recognize both `pl7.app/clusterId` and `pl7.app/vdj/clusterId` axis names (clonotype-clustering rename).
+  - Cluster-size query uses a namePattern matching both `pl7.app/clustering/clusterSize` and `pl7.app/vdj/clustering/clusterSize`.
+
 ## 4.0.0
 
 ### Major Changes

package/dist/tengo/lib/utils.lib.tengo

@@ -1,5 +1,6 @@
 
 
+ll := import("@platforma-sdk/workflow-tengo:ll")
 slices := import("@platforma-sdk/workflow-tengo:slices")
 json := import("json")
 
@@ -13,6 +14,13 @@ inVivoScoreSourceColumns := {
 
 
 
+isClusterIdAxis := func(name) {
+    return name == "pl7.app/clusterId" || name == "pl7.app/vdj/clusterId"
+}
+
+
+
+
 
 
 
@@ -61,7 +69,7 @@ findMatchingLinkerIndex := func(colsSpec, linkerColumns) {
 
     rankingClusterIdAxis := undefined
     for axis in colsSpec.axesSpec {
-        if axis.name == "pl7.app/clusterId" {
+        if isClusterIdAxis(axis.name) {
             rankingClusterIdAxis = axis
             break
         }
@@ -76,7 +84,7 @@ findMatchingLinkerIndex := func(colsSpec, linkerColumns) {
 
         linkerClusterIdAxis := undefined
         for axis in linkerCol.spec.axesSpec {
-            if axis.name == "pl7.app/clusterId" {
+            if isClusterIdAxis(axis.name) {
                 linkerClusterIdAxis = axis
                 break
             }
@@ -195,7 +203,7 @@ resolveClusterColumnHeader := func(args, columns, sortedLinkers) {
 
     selectedClusterIdAxis := undefined
     for axis in selectedLinkerSpec.axesSpec {
-        if axis.name == "pl7.app/clusterId" {
+        if isClusterIdAxis(axis.name) {
             selectedClusterIdAxis = axis
             break
         }
@@ -209,7 +217,7 @@ resolveClusterColumnHeader := func(args, columns, sortedLinkers) {
     for linkerIdx, col in sortedLinkers {
 
         for axis in col.spec.axesSpec {
-            if axis.name == "pl7.app/clusterId" {
+            if isClusterIdAxis(axis.name) {
 
                 if clusterAxisDomainsMatch(selectedClusterIdAxis, axis) {
                     return "clusterAxis_" + string(linkerIdx) + "_0"
@@ -409,7 +417,7 @@ initializeCloneTable := func(pframes, columns, args, datasetSpec) {
             addedCols = true
         }
 
-        if !is_undefined(clusterIdAxis) && clusterIdAxis.name == "pl7.app/clusterId" {
+        if !is_undefined(clusterIdAxis) && isClusterIdAxis(clusterIdAxis.name) {
             linkerClusterIdAxesWithIdx = append(linkerClusterIdAxesWithIdx, {
                 axis: clusterIdAxis,
                 linkerIdx: linkerIdx
@@ -423,7 +431,7 @@ initializeCloneTable := func(pframes, columns, args, datasetSpec) {
 
             clusterSizeClusterIdAxis := undefined
             for axis in col.spec.axesSpec {
-                if axis.name == "pl7.app/clusterId" {
+                if isClusterIdAxis(axis.name) {
                     clusterSizeClusterIdAxis = axis
                     break
                 }
@@ -459,11 +467,18 @@ initializeCloneTable := func(pframes, columns, args, datasetSpec) {
     }
 
 
+
     if !addedCols {
         cdr3Sequences := columns.getColumns("cdr3Sequences")
         if len(cdr3Sequences) > 0 {
-            cloneTable.add(cdr3Sequences[0], {header: "cdr3_fallback"})
+            cloneTable.add(cdr3Sequences[0], {header: "sequence_fallback"})
             addedCols = true
+        } else {
+            peptideMainSeqs := columns.getColumns("peptideMainSeqs")
+            if len(peptideMainSeqs) > 0 {
+                cloneTable.add(peptideMainSeqs[0], {header: "sequence_fallback"})
+                addedCols = true
+            }
         }
     }
 
@@ -594,11 +609,28 @@ detectBulkChain := func(seqCols) {
 
 
 
-initializeAssemSeqTable := func(pframes, columns, datasetSpec, isSingleCell) {
+
+initializeAssemSeqTable := func(pframes, columns, datasetSpec, isSingleCell, isScFv) {
     assemSeqTable := pframes.parquetFileBuilder()
     assemSeqTable.setAxisHeader(datasetSpec.axesSpec[1].name, "clonotypeKey")
 
-    seqCols := columns.getColumns("assemblingAaSeqs")
+    seqCols := undefined
+    if isScFv {
+        allowedFeatures := { "VDJRegion": true, "VDJRegionInFrame": true }
+        byChain := {}
+        for col in columns.getColumns("scFvPerChainSeqs") {
+            if is_undefined(allowedFeatures[col.spec.domain["pl7.app/vdj/feature"]]) { continue }
+            if col.spec.domain["pl7.app/vdj/scClonotypeChain/index"] != "primary" { continue }
+            sc := col.spec.domain["pl7.app/vdj/scClonotypeChain"]
+            if sc != "A" && sc != "B" { continue }
+            if is_undefined(byChain[sc]) { byChain[sc] = col }
+        }
+        seqCols = []
+        for _, c in byChain { seqCols = append(seqCols, c) }
+    } else {
+        seqCols = columns.getColumns("assemblingAaSeqs")
+    }
+
     for col in seqCols {
         headerName := makeHeaderName(col, "assemblingFeature", isSingleCell)
         assemSeqTable.add(col, {header: headerName})
@@ -606,7 +638,7 @@ initializeAssemSeqTable := func(pframes, columns, datasetSpec, isSingleCell) {
 
     assemSeqTable.mem("16GiB")
     assemSeqTable.cpu(1)
-    
+
 
     bulkChain := undefined
     if !isSingleCell {

package/package.json

@@ -1,16 +1,16 @@
 {
   "name": "@platforma-open/milaboratories.top-antibodies.workflow",
-  "version": "4.0.0",
+  "version": "4.0.2",
   "type": "module",
   "description": "Block Workflow",
   "dependencies": {
     "@platforma-sdk/workflow-tengo": "5.20.0",
     "@platforma-open/milaboratories.software-anarci": "^0.0.3",
     "@platforma-open/milaboratories.top-antibodies.sample-clonotypes": "2.1.2",
-    "@platforma-open/milaboratories.top-antibodies.assembling-fasta": "1.3.2",
-    "@platforma-open/milaboratories.top-antibodies.anarci-kabat": "1.4.3",
     "@platforma-open/milaboratories.top-antibodies.spectratype": "1.8.3",
-    "@platforma-open/milaboratories.top-antibodies.umap": "1.2.3"
+    "@platforma-open/milaboratories.top-antibodies.umap": "1.2.3",
+    "@platforma-open/milaboratories.top-antibodies.anarci-kabat": "1.4.3",
+    "@platforma-open/milaboratories.top-antibodies.assembling-fasta": "1.3.2"
   },
   "devDependencies": {
     "@platforma-sdk/tengo-builder": "2.5.20"

package/src/main.tpl.tengo

@@ -75,9 +75,11 @@ wf.prepare(func(args){
         bundleBuilder.addAnchor("selectedCluster", args.diversificationColumn)
     }
 
-    // Add cluster size columns from clustering blocks
+    // Add cluster size columns from clustering blocks. The namePattern matches
+    // both the post-peptide-adaptation name and the pre-peptide `pl7.app/vdj/`
+    // variant so older clonotype-clustering instances remain usable.
     bundleBuilder.addMulti({
-        name: "pl7.app/vdj/clustering/clusterSize",
+        namePattern: "^pl7\\.app/(vdj/)?clustering/clusterSize$",
         partialAxesMatch: true
     }, "clusterSizes")
 
@@ -109,12 +111,37 @@ wf.prepare(func(args){
 		}
 	}, "JGenes")
 
-	// Add assembling feature aminoacid sequences (bulk, sc, scFv)
+	// Add assembling feature aminoacid sequences (bulk, sc).
 	bundleBuilder.addMulti({
 		axes: [{ anchor: "main", idx: 1 }], // Clonotype axis
+		name: "pl7.app/vdj/sequence",
 		annotations: { "pl7.app/vdj/isAssemblingFeature": "true" },
 		domain: { "pl7.app/alphabet": "aminoacid" }
 	}, "assemblingAaSeqs")
+
+	// Detect scFv mode by presence of the combined construct column.
+	bundleBuilder.addMulti({
+		axes: [{ anchor: "main", idx: 1 }],
+		name: "pl7.app/vdj/scFv-sequence"
+	}, "scFvDetect")
+
+	// Per-chain VDJRegion AA sequences for ANARCI/Kabat in scFv mode.
+	bundleBuilder.addMulti({
+		axes: [{ anchor: "main", idx: 1 }],
+		name: "pl7.app/vdj/sequence",
+		domain: { "pl7.app/alphabet": "aminoacid" }
+	}, "scFvPerChainSeqs")
+
+	// Peptide main sequence — single-axis column on variantKey, used as
+	// modality-aware fallback when no filter/ranking columns load.
+	bundleBuilder.addMulti({
+		axes: [{ anchor: "main", idx: 1 }],
+		annotations: {
+			"pl7.app/isAssemblingFeature": "true",
+			"pl7.app/isMainSequence": "true"
+		},
+		domain: { "pl7.app/alphabet": "aminoacid" }
+	}, "peptideMainSeqs")
     
     return {
         columns: bundleBuilder.build()
@@ -224,7 +251,8 @@ wf.body(func(args) {
                 if args.kabatNumbering == true {
                     ////////// Assembling AA sequences //////////
                     // Initialize and build assembling sequence table
-                    assemInit := utils.initializeAssemSeqTable(pframes, columns, datasetSpec, isSingleCell)
+                    isScFv := len(columns.getColumns("scFvDetect")) > 0
+                    assemInit := utils.initializeAssemSeqTable(pframes, columns, datasetSpec, isSingleCell, isScFv)
                     assemSeqTableBuilt := assemInit.assemSeqTable
                     bulkChain := assemInit.bulkChain
                     seqCols := assemInit.seqCols

package/src/utils.lib.tengo

@@ -1,5 +1,6 @@
 // Utility functions for antibody-tcr-lead-selection workflow
 
+ll := import("@platforma-sdk/workflow-tengo:ll")
 slices := import("@platforma-sdk/workflow-tengo:slices")
 json := import("json")
 
@@ -10,6 +11,13 @@ inVivoScoreSourceColumns := {
     "nMutations": "pl7.app/vdj/sequence/nMutations"
 }
 
+// Cluster-id axis names — both unprefixed (post-peptide-adaptation) and
+// `pl7.app/vdj/`-prefixed (pre-peptide) so older clonotype-clustering
+// instances remain selectable as diversification linkers.
+isClusterIdAxis := func(name) {
+    return name == "pl7.app/clusterId" || name == "pl7.app/vdj/clusterId"
+}
+
 /**
  * Checks if two clusterId axes have matching domains.
  * Used to determine if two columns belong to the same clustering run.
@@ -61,7 +69,7 @@ findMatchingLinkerIndex := func(colsSpec, linkerColumns) {
     // Find the clusterId axis in the ranking column
     rankingClusterIdAxis := undefined
     for axis in colsSpec.axesSpec {
-        if axis.name == "pl7.app/clusterId" {
+        if isClusterIdAxis(axis.name) {
             rankingClusterIdAxis = axis
             break
         }
@@ -76,7 +84,7 @@ findMatchingLinkerIndex := func(colsSpec, linkerColumns) {
         // Get the clusterId axis from the linker column
         linkerClusterIdAxis := undefined
         for axis in linkerCol.spec.axesSpec {
-            if axis.name == "pl7.app/clusterId" {
+            if isClusterIdAxis(axis.name) {
                 linkerClusterIdAxis = axis
                 break
             }
@@ -195,7 +203,7 @@ resolveClusterColumnHeader := func(args, columns, sortedLinkers) {
     // Find the clusterId axis in the selected linker
     selectedClusterIdAxis := undefined
     for axis in selectedLinkerSpec.axesSpec {
-        if axis.name == "pl7.app/clusterId" {
+        if isClusterIdAxis(axis.name) {
             selectedClusterIdAxis = axis
             break
         }
@@ -209,7 +217,7 @@ resolveClusterColumnHeader := func(args, columns, sortedLinkers) {
     for linkerIdx, col in sortedLinkers {
         // Get the clusterId axis from this linker
         for axis in col.spec.axesSpec {
-            if axis.name == "pl7.app/clusterId" {
+            if isClusterIdAxis(axis.name) {
                 // Use clusterAxisDomainsMatch for proper domain comparison
                 if clusterAxisDomainsMatch(selectedClusterIdAxis, axis) {
                     return "clusterAxis_" + string(linkerIdx) + "_0"
@@ -409,7 +417,7 @@ initializeCloneTable := func(pframes, columns, args, datasetSpec) {
             addedCols = true
         }
         // Collect clusterId axes from linker columns to match cluster size columns
-        if !is_undefined(clusterIdAxis) && clusterIdAxis.name == "pl7.app/clusterId" {
+        if !is_undefined(clusterIdAxis) && isClusterIdAxis(clusterIdAxis.name) {
             linkerClusterIdAxesWithIdx = append(linkerClusterIdAxesWithIdx, {
                 axis: clusterIdAxis,
                 linkerIdx: linkerIdx
@@ -423,7 +431,7 @@ initializeCloneTable := func(pframes, columns, args, datasetSpec) {
             // Find the clusterId axis in this cluster size column
             clusterSizeClusterIdAxis := undefined
             for axis in col.spec.axesSpec {
-                if axis.name == "pl7.app/clusterId" {
+                if isClusterIdAxis(axis.name) {
                     clusterSizeClusterIdAxis = axis
                     break
                 }
@@ -458,12 +466,19 @@ initializeCloneTable := func(pframes, columns, args, datasetSpec) {
         }
     }
 
-    // Fallback: if no columns added, add at least one CDR3 sequence column
+    // Fallback: if no columns added, add a single-axis trunk-keyed sequence
+    // column. Try VDJ CDR3 first
     if !addedCols {
         cdr3Sequences := columns.getColumns("cdr3Sequences")
         if len(cdr3Sequences) > 0 {
-            cloneTable.add(cdr3Sequences[0], {header: "cdr3_fallback"})
+            cloneTable.add(cdr3Sequences[0], {header: "sequence_fallback"})
             addedCols = true
+        } else {
+            peptideMainSeqs := columns.getColumns("peptideMainSeqs")
+            if len(peptideMainSeqs) > 0 {
+                cloneTable.add(peptideMainSeqs[0], {header: "sequence_fallback"})
+                addedCols = true
+            }
         }
     }
 
@@ -587,18 +602,35 @@ detectBulkChain := func(seqCols) {
 
 /**
  * Initializes and builds assembling sequence table with assembling AA sequences.
- * 
+
  * @param pframes - PFrames import
  * @param columns - PBundle containing all columns
  * @param datasetSpec - Dataset specification with axes
  * @param isSingleCell - Whether the data is single cell
+ * @param isScFv - Whether the input is scFv data
  * @return Map with keys: assemSeqTable (built table), bulkChain, seqCols
  */
-initializeAssemSeqTable := func(pframes, columns, datasetSpec, isSingleCell) {
+initializeAssemSeqTable := func(pframes, columns, datasetSpec, isSingleCell, isScFv) {
     assemSeqTable := pframes.parquetFileBuilder()
     assemSeqTable.setAxisHeader(datasetSpec.axesSpec[1].name, "clonotypeKey")
 
-    seqCols := columns.getColumns("assemblingAaSeqs")
+    seqCols := undefined
+    if isScFv {
+        allowedFeatures := { "VDJRegion": true, "VDJRegionInFrame": true }
+        byChain := {}
+        for col in columns.getColumns("scFvPerChainSeqs") {
+            if is_undefined(allowedFeatures[col.spec.domain["pl7.app/vdj/feature"]]) { continue }
+            if col.spec.domain["pl7.app/vdj/scClonotypeChain/index"] != "primary" { continue }
+            sc := col.spec.domain["pl7.app/vdj/scClonotypeChain"]
+            if sc != "A" && sc != "B" { continue }
+            if is_undefined(byChain[sc]) { byChain[sc] = col }
+        }
+        seqCols = []
+        for _, c in byChain { seqCols = append(seqCols, c) }
+    } else {
+        seqCols = columns.getColumns("assemblingAaSeqs")
+    }
+
     for col in seqCols {
         headerName := makeHeaderName(col, "assemblingFeature", isSingleCell)
         assemSeqTable.add(col, {header: headerName})
@@ -606,7 +638,7 @@ initializeAssemSeqTable := func(pframes, columns, datasetSpec, isSingleCell) {
 
     assemSeqTable.mem("16GiB")
     assemSeqTable.cpu(1)
-    
+
     // Detect bulk chain if needed
     bulkChain := undefined
     if !isSingleCell {