@platforma-open/milaboratories.mixcr-amplicon-alignment.workflow 1.19.1 → 1.19.3

package/.turbo/turbo-build.log

@@ -1,6 +1,6 @@
  WARN  Issue while reading "/home/runner/work/mixcr-amplicon-alignment/mixcr-amplicon-alignment/.npmrc". Failed to replace env in config: ${NPMJS_TOKEN}
 
-> @platforma-open/milaboratories.mixcr-amplicon-alignment.workflow@1.19.1 build /home/runner/work/mixcr-amplicon-alignment/mixcr-amplicon-alignment/workflow
+> @platforma-open/milaboratories.mixcr-amplicon-alignment.workflow@1.19.3 build /home/runner/work/mixcr-amplicon-alignment/mixcr-amplicon-alignment/workflow
 > rm -rf dist && pl-tengo check && pl-tengo build
 
 Processing "src/aggregate-by-clonotype-key.tpl.tengo"...

package/CHANGELOG.md

@@ -1,5 +1,17 @@
 # @platforma-open/milaboratories.mixcr-amplicon-alignment.workflow
 
+## 1.19.3
+
+### Patch Changes
+
+- 8685e8d: Add new mutation columns
+
+## 1.19.2
+
+### Patch Changes
+
+- 4aeac00: Link memory limits of downstream operations (XSV conversion, MiXCR export, PTabler) to user-specified perProcessMemGB override to prevent over-scheduling on local runs
+
 ## 1.19.1
 
 ### Patch Changes

package/dist/tengo/lib/calculate-export-specs.lib.tengo

@@ -619,7 +619,74 @@ inFrameFeatures := {
 		exportArgs += [ [ "-allAAMutationsCount", "FR1Begin", "FR4End" ] ]
 		
 	}
-	
+
+
+	if assemblingFeature == "VDJRegion" {
+		orderP = 7500
+		nMutationsFeature := coreGeneFeatures["V"] + "," + coreGeneFeatures["J"]
+
+		crossRegionMutations := [ {
+			exportFlag: "nMutationsCount",
+			feature: nMutationsFeature,
+			id: "n-mutations-total",
+			label: "Nt mutations",
+			specName: "pl7.app/vdj/sequence/nMutations",
+			rankingOrder: "decreasing"
+		}, {
+			exportFlag: "aaMutationsCount",
+			feature: "CDR1,CDR2",
+			id: "aa-mutations-cdr",
+			label: "AA mutations (CDR)",
+			specName: "pl7.app/vdj/sequence/nAAMutationsCDR",
+			rankingOrder: "decreasing"
+		}, {
+			exportFlag: "aaMutationsCount",
+			feature: "FR1,FR2,FR3,FR4",
+			id: "aa-mutations-fwr",
+			label: "AA mutations (FWR)",
+			specName: "pl7.app/vdj/sequence/nAAMutationsFWR",
+			rankingOrder: "increasing"
+		} ]
+
+		for col in crossRegionMutations {
+			columnsSpecPerClonotypeNoAggregates += [ {
+				column: col.exportFlag + col.feature,
+				id: col.id,
+				allowNA: true,
+				naRegex: "region_not_covered",
+				spec: {
+					valueType: "Int",
+					name: col.specName,
+					annotations: a(orderP, false, {
+						"pl7.app/label": col.label,
+						"pl7.app/isScore": "true",
+						"pl7.app/score/rankingOrder": col.rankingOrder
+					})
+				}
+			} ]
+			exportArgs += [ [ "-" + col.exportFlag, col.feature ] ]
+			orderP -= 100
+		}
+
+
+		columnsSpecPerClonotypeNoAggregates += [ {
+			column: "fractionCDRMutations",
+			id: "fraction-cdr-mutations",
+			naRegex: "^[a-z_]*$",
+			allowNA: true,
+			spec: {
+				valueType: "Double",
+				name: "pl7.app/vdj/sequence/fractionCDRMutations",
+				annotations: a(orderP, false, {
+					"pl7.app/label": "CDR mutation fraction",
+					"pl7.app/isScore": "true",
+					"pl7.app/score/rankingOrder": "decreasing",
+					"pl7.app/format": ".2f"
+				})
+			}
+		} ]
+	}
+
 
 
 	germlineVFeature := "GermlineVCDR3Part"

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@platforma-open/milaboratories.mixcr-amplicon-alignment.workflow",
-  "version": "1.19.1",
+  "version": "1.19.3",
   "description": "MiXCR Amplicon Alignment Workflow",
   "type": "module",
   "dependencies": {

package/src/aggregate-by-clonotype-key.tpl.tengo

@@ -72,11 +72,16 @@ self.body(func(inputs) {
 	}
 
 	aggExpressions := []
+	hasFractionCDRMutations := false
 
 	for colDef in schemaPerClonotypeNoAggregates {
 		if colDef.column == "clonotypeLabel" {
 			continue
 		}
+		if colDef.column == "fractionCDRMutations" {
+			hasFractionCDRMutations = true
+			continue
+		}
 		aggExpressions = append(aggExpressions,
 			pt.col(colDef.column).maxBy(pt.col(mainAbundanceColumnNormalized)).alias(colDef.column)
 		)
@@ -90,6 +95,24 @@ self.body(func(inputs) {
 
 	aggregatedDf := currentDf.groupBy("clonotypeKey").agg(aggExpressions...)
 
+	// Calculate CDR mutation fraction: CDR / (CDR + FWR), fallback 1.0 when NA or zero denominator
+	if hasFractionCDRMutations {
+		cdr := "aaMutationsCountCDR1,CDR2"
+		fwr := "aaMutationsCountFR1,FR2,FR3,FR4"
+		aggregatedDf = aggregatedDf.withColumns(
+			pt.when(
+				pt.col(cdr).isNotNull().
+					and(pt.col(fwr).isNotNull()).
+					and(pt.col(cdr).cast("Double").plus(pt.col(fwr).cast("Double")).gt(0.0))
+			).then(
+				pt.col(cdr).cast("Double").truediv(
+					pt.col(cdr).cast("Double").plus(pt.col(fwr).cast("Double"))
+				)
+			).otherwise(pt.lit(1.0)).
+				alias("fractionCDRMutations")
+		)
+	}
+
 	aggregatedDf = clonotypeLabel.addClonotypeLabelColumnsPt(aggregatedDf, "clonotypeKey", "clonotypeLabel", pt)
 
 	cdr3Df := aggregatedDf.select(

package/src/calculate-export-specs.lib.tengo

@@ -619,7 +619,74 @@ inFrameFeatures := {
 		exportArgs += [ [ "-allAAMutationsCount", "FR1Begin", "FR4End" ] ]
 		
 	}
-	
+
+	// Cross-region SHM mutation counts (only for full VDJRegion assembling)
+	if assemblingFeature == "VDJRegion" {
+		orderP = 7500
+		nMutationsFeature := coreGeneFeatures["V"] + "," + coreGeneFeatures["J"]
+
+		crossRegionMutations := [ {
+			exportFlag: "nMutationsCount",
+			feature: nMutationsFeature,
+			id: "n-mutations-total",
+			label: "Nt mutations",
+			specName: "pl7.app/vdj/sequence/nMutations",
+			rankingOrder: "decreasing"
+		}, {
+			exportFlag: "aaMutationsCount",
+			feature: "CDR1,CDR2",
+			id: "aa-mutations-cdr",
+			label: "AA mutations (CDR)",
+			specName: "pl7.app/vdj/sequence/nAAMutationsCDR",
+			rankingOrder: "decreasing"
+		}, {
+			exportFlag: "aaMutationsCount",
+			feature: "FR1,FR2,FR3,FR4",
+			id: "aa-mutations-fwr",
+			label: "AA mutations (FWR)",
+			specName: "pl7.app/vdj/sequence/nAAMutationsFWR",
+			rankingOrder: "increasing"
+		} ]
+
+		for col in crossRegionMutations {
+			columnsSpecPerClonotypeNoAggregates += [ {
+				column: col.exportFlag + col.feature,
+				id: col.id,
+				allowNA: true,
+				naRegex: "region_not_covered",
+				spec: {
+					valueType: "Int",
+					name: col.specName,
+					annotations: a(orderP, false, {
+						"pl7.app/label": col.label,
+						"pl7.app/isScore": "true",
+						"pl7.app/score/rankingOrder": col.rankingOrder
+					})
+				}
+			} ]
+			exportArgs += [ [ "-" + col.exportFlag, col.feature ] ]
+			orderP -= 100
+		}
+
+		// CDR mutation fraction (computed in aggregate-by-clonotype-key)
+		columnsSpecPerClonotypeNoAggregates += [ {
+			column: "fractionCDRMutations",
+			id: "fraction-cdr-mutations",
+			naRegex: "^[a-z_]*$",
+			allowNA: true,
+			spec: {
+				valueType: "Double",
+				name: "pl7.app/vdj/sequence/fractionCDRMutations",
+				annotations: a(orderP, false, {
+					"pl7.app/label": "CDR mutation fraction",
+					"pl7.app/isScore": "true",
+					"pl7.app/score/rankingOrder": "decreasing",
+					"pl7.app/format": ".2f"
+				})
+			}
+		} ]
+	}
+
 	// Export germline CDR3 part mutations for both VDJRegion and CDR3 assembling features
 	// These will be summed to create the CDR3 mutations count columns
 	germlineVFeature := "GermlineVCDR3Part"

package/src/mixcr-export.tpl.tengo

@@ -8,6 +8,7 @@ exec := import("@platforma-sdk/workflow-tengo:exec")
 pt := import("@platforma-sdk/workflow-tengo:pt")
 clonotypeLabel := import(":clonotype-label")
 
+math := import("math")
 json := import("json")
 text := import("text")
 
@@ -176,6 +177,15 @@ self.body(func(inputs) {
 
 	useProductiveFilter := is_undefined(stopCodonTypes) || len(stopCodonTypes) == 0
 
+	// Memory for downstream operations, linked to user override with hardcoded floors
+	baseMemGB := 64
+	if !is_undefined(params.perProcessMemGB) {
+		baseMemGB = params.perProcessMemGB
+	}
+	memGB := func(floorGB, divisor) {
+		return string(int(math.max(floorGB, baseMemGB / divisor))) + "GB"
+	}
+
 	hashKeyDerivationExpressionPt := func(sourceColumns) {
 		return pt.concatStr(
 			slices.map(sourceColumns, func(colName) { return pt.col(colName) }),
@@ -186,7 +196,7 @@ self.body(func(inputs) {
 	createExport := func(additionalAction) {
 		mixcrCmdBuilder := exec.builder().
 			inMediumQueue().
-			mem("12GB").
+			mem(memGB(12, 4)).
 		    cpu(2).
 			printErrStreamToStdout().
 			software(mixcrSw).
@@ -233,7 +243,7 @@ self.body(func(inputs) {
 	// Simplified PTabler processing for main TSV output
 	wfMain := pt.workflow().
 		inMediumQueue().
-		mem("8GB").
+		mem(memGB(8, 4)).
 		cpu(2)
 		
 	frameInputMap := {

package/src/process.tpl.tengo

@@ -9,6 +9,7 @@ pframes := import("@platforma-sdk/workflow-tengo:pframes")
 slices := import("@platforma-sdk/workflow-tengo:slices")
 maps := import("@platforma-sdk/workflow-tengo:maps")
 
+math := import("math")
 json := import("json")
 text := import("text")
 
@@ -71,6 +72,18 @@ self.body(func(inputs) {
 	limitInput := inputs.limitInput
 	perProcessMemGB := params.perProcessMemGB
 	perProcessCPUs := params.perProcessCPUs
+
+	// Base memory for scheduling downstream operations, linked to user override
+	// with current hardcoded values as floors (see memGB helper below)
+	baseMemGB := 64
+	if !is_undefined(perProcessMemGB) {
+		baseMemGB = perProcessMemGB
+	}
+	// Returns max(floor, baseMemGB / divisor) as a string like "32GB"
+	memGB := func(floorGB, divisor) {
+		return string(int(math.max(floorGB, baseMemGB / divisor))) + "GB"
+	}
+
 	fileExtension := inputSpec.domain["pl7.app/fileExtension"]
 	sampleIdAxisSpec := inputSpec.axesSpec[0]
 
@@ -249,7 +262,7 @@ self.body(func(inputs) {
 			storageFormat: "Parquet",
 			partitionKeyLength: 0
 		},
-		mem: "16GB",
+		mem: memGB(16, 2),
 		cpu: 2,
 		name: "byCloneKeyBySample",
 		path: ["tsv"]
@@ -271,7 +284,8 @@ self.body(func(inputs) {
 					aminoAcidSeqColumnPairs: aminoAcidSeqColumnPairs,
 					cdr3SeqColumns: cdr3SeqColumns,
 					stopCodonTypes: params.stopCodonTypes,
-					stopCodonReplacements: params.stopCodonReplacements
+					stopCodonReplacements: params.stopCodonReplacements,
+					perProcessMemGB: perProcessMemGB
 				}, { removeUndefs: true })
 			}
 		}
@@ -302,7 +316,7 @@ self.body(func(inputs) {
 			columns: columnsSpecPerClonotypeNoAggregates + columnsSpecPerClonotypeAggregates,
 			storageFormat: "Parquet"
 		},
-		mem: "12GB",
+		mem: memGB(12, 4),
 		cpu: 2,
 		name: "aggregates",
 		path: ["tsv"]
@@ -314,7 +328,7 @@ self.body(func(inputs) {
 			columns: cdr3DistanceColumnsSpec,
 			storageFormat: "Parquet"
 		},
-		mem: "8GB",
+		mem: memGB(8, 8),
 		cpu: 1,
 		name: "cdr3Distances",
 		path: ["cdr3DistancesTsv"]