Merge branch 'hotfix/v2.1.12'

CHANGELOG

@@ -1,4 +1,12 @@
 
+MiXCR 2.1.12 (22 Aug 2018)
+========================
+
+-- Fixes rare inconsistency in behaviour between paired-end and single-end aligners, clone 
+ assembler and partial sequence assembler in respect to target D gene set selection based on the 
+ chains on V, J and C genes (#408)
+
+
 MiXCR 2.1.11 (15 Jun 2018)
 ========================
 

pom.xml

@@ -33,7 +33,7 @@
 
  <groupId>com.milaboratory</groupId>
  <artifactId>mixcr</artifactId>
- <version>2.1.11</version>
+ <version>2.1.12</version>
  <packaging>jar</packaging>
  <name>MiXCR</name>
 

src/main/java/com/milaboratory/mixcr/assembler/CloneFactory.java

@@ -229,7 +229,8 @@ Clone create(int id, CloneAccumulator accumulator) {
  if (from < to)
  hits.put(GeneType.Diversity,
  dAligner.align(sequenceToAlign,
- VDJCAligner.getPossibleDLoci(hits.get(GeneType.Variable), hits.get(GeneType.Joining)),
+ VDJCAligner.getPossibleDLoci(hits.get(GeneType.Variable), hits.get(GeneType.Joining),
+ null),
  from, to, indexOfAssemblingFeatureWithD,
  assemblingFeatures.length));
  else

src/main/java/com/milaboratory/mixcr/assembler/CloneFactoryParameters.java

@@ -32,6 +32,7 @@
 
 import com.fasterxml.jackson.annotation.JsonAutoDetect;
 import com.fasterxml.jackson.annotation.JsonCreator;
+import com.fasterxml.jackson.annotation.JsonInclude;
 import com.fasterxml.jackson.annotation.JsonProperty;
 import com.fasterxml.jackson.databind.annotation.JsonSerialize;
 import com.milaboratory.mixcr.basictypes.ClonalUpdatableParameters;
@@ -93,25 +94,25 @@ public VJCClonalAlignerParameters getVJCParameters(GeneType geneType) {
  }
 
  @JsonProperty("vParameters")
- @JsonSerialize(include = JsonSerialize.Inclusion.NON_NULL)
+ @JsonInclude(JsonInclude.Include.NON_NULL)
  public VJCClonalAlignerParameters getVParameters() {
  return vdcParameters.get(GeneType.Variable);
  }
 
  @JsonProperty("jParameters")
- @JsonSerialize(include = JsonSerialize.Inclusion.NON_NULL)
+ @JsonInclude(JsonInclude.Include.NON_NULL)
  public VJCClonalAlignerParameters getJParameters() {
  return vdcParameters.get(GeneType.Joining);
  }
 
  @JsonProperty("cParameters")
- @JsonSerialize(include = JsonSerialize.Inclusion.NON_NULL)
+ @JsonInclude(JsonInclude.Include.NON_NULL)
  public VJCClonalAlignerParameters getCParameters() {
  return vdcParameters.get(GeneType.Constant);
  }
 
  @JsonProperty("dParameters")
- @JsonSerialize(include = JsonSerialize.Inclusion.NON_NULL)
+ @JsonInclude(JsonInclude.Include.NON_NULL)
  public DClonalAlignerParameters getDParameters() {
  return dParameters;
  }

src/main/java/com/milaboratory/mixcr/partialassembler/PartialAlignmentsAssemblerAligner.java

@@ -297,7 +297,7 @@ protected VDJCAlignmentResult<VDJCMultiRead> process0(VDJCMultiRead input) {
  if (vAl == null || jAl == null || singleDAligner == null)
  dResult = new VDJCHit[0];
  else
- dResult = singleDAligner.align(targets[dGeneTarget].getSequence(), getPossibleDLoci(vResult, jResult),
+ dResult = singleDAligner.align(targets[dGeneTarget].getSequence(), getPossibleDLoci(vResult, jResult, null),
  vAl.getSequence2Range().getTo(),
  jAl.getSequence2Range().getFrom(),
  dGeneTarget, nReads);

src/main/java/com/milaboratory/mixcr/vdjaligners/VDJCAligner.java

@@ -30,10 +30,11 @@
 package com.milaboratory.mixcr.vdjaligners;
 
 import cc.redberry.pipe.Processor;
+import com.milaboratory.core.alignment.batch.AlignmentHit;
 import com.milaboratory.core.io.sequence.SequenceRead;
 import com.milaboratory.core.io.sequence.SingleRead;
+import com.milaboratory.mixcr.basictypes.HasGene;
 import com.milaboratory.mixcr.basictypes.VDJCAlignments;
-import com.milaboratory.mixcr.basictypes.VDJCHit;
 import com.milaboratory.util.HashFunctions;
 import com.milaboratory.util.RandomUtil;
 import io.repseq.core.Chains;
@@ -172,12 +173,81 @@ public static VDJCAligner createAligner(VDJCAlignerParameters alignerParameters,
  : new VDJCAlignerS(alignerParameters);
  }
 
- public static Chains getPossibleDLoci(VDJCHit[] vHits, VDJCHit[] jHits) {
- Chains chains = new Chains();
- for (VDJCHit h : vHits)
- chains = chains.merge(h.getGene().getChains());
- for (VDJCHit h : jHits)
- chains = chains.merge(h.getGene().getChains());
+ // public static Chains getPossibleDLoci(VDJCHit[] vHits, VDJCHit[] jHits) {
+ // Chains chains = new Chains();
+ // for (VDJCHit h : vHits)
+ // chains = chains.merge(h.getGene().getChains());
+ // for (VDJCHit h : jHits)
+ // chains = chains.merge(h.getGene().getChains());
+ // return chains;
+ // }
+
+ /*
+ * Common utility methods
+ */
+
+ static Chains getVJCommonChains(final HasGene[] vHits, final HasGene[] jHits) {
+ return getChains(vHits).intersection(getChains(jHits));
+ }
+
+ /**
+ * Merge chains of all V/D/J/C hits.
+ *
+ * Allows null as input.
+ */
+ public static Chains getChains(final HasGene[] hits) {
+ if (hits == null || hits.length == 0)
+ return Chains.ALL;
+ Chains chains = hits[0].getGene().getChains();
+ for (int i = 1; i < hits.length; i++)
+ chains = chains.merge(hits[i].getGene().getChains());
  return chains;
  }
+
+ /**
+ * Calculates possible chains for D gene alignment, in the presence of following V, J and C genes.
+ *
+ * Allows nulls as input.
+ */
+ public static Chains getPossibleDLoci(HasGene[] vHits, HasGene[] jHits, HasGene[] cHits) {
+ Chains intersection = getChains(vHits)
+ .intersection(getChains(jHits))
+ .intersection(getChains(cHits));
+
+ if (!intersection.isEmpty())
+ return intersection;
+
+ // If intersection is empty, we are working with chimera
+ // lets calculate all possible D loci by merging
+ Chains chains = Chains.EMPTY;
+ if (vHits != null)
+ for (HasGene h : vHits)
+ chains = chains.merge(h.getGene().getChains());
+ if (jHits != null)
+ for (HasGene h : jHits)
+ chains = chains.merge(h.getGene().getChains());
+ if (cHits != null)
+ for (HasGene h : cHits)
+ chains = chains.merge(h.getGene().getChains());
+
+ return chains;
+ }
+
+ static HasGene[] wrapAlignmentHits(final AlignmentHit<?, VDJCGene>[] hits) {
+ if (hits == null)
+ return null;
+ HasGene[] res = new HasGene[hits.length];
+ for (int i = 0; i < hits.length; i++)
+ res[i] = wrapAlignmentHit(hits[i]);
+ return res;
+ }
+
+ static HasGene wrapAlignmentHit(final AlignmentHit<?, VDJCGene> hit) {
+ return new HasGene() {
+ @Override
+ public VDJCGene getGene() {
+ return hit.getRecordPayload();
+ }
+ };
+ }
 }

src/main/java/com/milaboratory/mixcr/vdjaligners/VDJCAlignerAbstract.java

@@ -60,14 +60,14 @@ public VDJCAlignerAbstract(VDJCAlignerParameters parameters) {
  }
 
  VDJCAlignerAbstract(boolean initialized,
-  VDJCAlignerParameters parameters,
-  EnumMap<GeneType, List<VDJCGene>> genesToAlign,
-  List<VDJCGene> usedGenes,
-  SingleDAligner singleDAligner,
-  EnumMap<GeneType, HashMap<String, BitArray>> filters,
-  BatchAlignerWithBaseWithFilter<NucleotideSequence, VDJCGene, AlignmentHit<NucleotideSequence, VDJCGene>> vAligner,
-  BatchAlignerWithBaseWithFilter<NucleotideSequence, VDJCGene, AlignmentHit<NucleotideSequence, VDJCGene>> jAligner,
-  BatchAlignerWithBaseWithFilter<NucleotideSequence, VDJCGene, AlignmentHit<NucleotideSequence, VDJCGene>> cAligner) {
+ VDJCAlignerParameters parameters,
+ EnumMap<GeneType, List<VDJCGene>> genesToAlign,
+ List<VDJCGene> usedGenes,
+ SingleDAligner singleDAligner,
+ EnumMap<GeneType, HashMap<String, BitArray>> filters,
+ BatchAlignerWithBaseWithFilter<NucleotideSequence, VDJCGene, AlignmentHit<NucleotideSequence, VDJCGene>> vAligner,
+ BatchAlignerWithBaseWithFilter<NucleotideSequence, VDJCGene, AlignmentHit<NucleotideSequence, VDJCGene>> jAligner,
+ BatchAlignerWithBaseWithFilter<NucleotideSequence, VDJCGene, AlignmentHit<NucleotideSequence, VDJCGene>> cAligner) {
  super(initialized, parameters, genesToAlign, usedGenes);
  this.singleDAligner = singleDAligner;
  this.filters = filters;

src/main/java/com/milaboratory/mixcr/vdjaligners/VDJCAlignerPVFirst.java

@@ -366,7 +366,7 @@ PAlignmentHelper alignVThenJ(final Target target) {
 
  boolean vChimera = checkAndEliminateChimera(vAl1, vAl2, GeneType.Variable);
 
- PairedHit[] vHits = extractDoubleHits(vAl1, vAl2);
+ PairedHit[] vHits = createPairedHits(vAl1, vAl2);
 
  /*
  * Step 2: of round of V gene alignment with more precise algorithm
@@ -433,7 +433,7 @@ PAlignmentHelper alignVThenJ(final Target target) {
 
  boolean jChimera = checkAndEliminateChimera(jAl1, jAl2, GeneType.Joining);
 
- PairedHit[] jHits = extractDoubleHits(jAl1, jAl2);
+ PairedHit[] jHits = createPairedHits(jAl1, jAl2);
 
  for (PairedHit jHit : jHits) {
  if (jHit.hit0 == null) {
@@ -584,7 +584,7 @@ List<AlignmentHit<NucleotideSequence, VDJCGene>> performJAlignment(
  /**
  * Converts two AlignmentResults to an array of paired hits (each paired hit for a particular V of J gene)
  */
- static PairedHit[] extractDoubleHits(List<AlignmentHit<NucleotideSequence, VDJCGene>>... results) {
+ static PairedHit[] createPairedHits(List<AlignmentHit<NucleotideSequence, VDJCGene>>... results) {
  Map<VDJCGeneId, PairedHit> hits = new HashMap<>();
  addHits(hits, results[0], 0);
  addHits(hits, results[1], 1);
@@ -612,28 +612,6 @@ static void addHits(Map<VDJCGeneId, PairedHit> hits,
  }
  }
 
- static Chains getVJCommonChains(final PairedHit[] vHits, final PairedHit[] jHits) {
- return getChains(vHits).intersection(getChains(jHits));
- }
-
- static Chains getChains(final PairedHit[] hits) {
- if (hits.length == 0)
- return Chains.ALL;
- Chains chains = hits[0].getGene().getChains();
- for (int i = 1; i < hits.length; i++)
- chains = chains.merge(hits[i].getGene().getChains());
- return chains;
- }
-
- static Chains getChains(final VDJCHit[] hits) {
- if (hits.length == 0)
- return Chains.ALL;
- Chains chains = hits[0].getGene().getChains();
- for (int i = 1; i < hits.length; i++)
- chains = chains.merge(hits[i].getGene().getChains());
- return chains;
- }
-
  static final PreVDJCHit[] zeroArray = new PreVDJCHit[0];
  @SuppressWarnings("unchecked")
  static final AlignmentHit<NucleotideSequence, VDJCGene>[] zeroKArray = new AlignmentHit[0];
@@ -819,39 +797,6 @@ public void filterHits(float minTotalScore, int maxHits) {
  }
  }
 
- public static Chains getPossibleDLoci(PairedHit[] vHits, PairedHit[] jHits, VDJCHit[] cHits) {
- Chains intersection = getChains(vHits)
- .intersection(getChains(jHits))
- .intersection(getChains(cHits));
-
- if (!intersection.isEmpty())
- return intersection;
-
- // If intersection is empty, we are working with chimer
- // lets calculate all possible D loci
- Chains chains = new Chains();
- for (PairedHit h : vHits)
- chains = chains.merge(h.getGene().getChains());
- for (PairedHit h : jHits)
- chains = chains.merge(h.getGene().getChains());
- for (VDJCHit h : cHits)
- chains = chains.merge(h.getGene().getChains());
-
- return chains;
- }
-
- ///**
- // * Converts array of "internal" PairedHits to a double array of KAlignmentHits to pass this value to a VDJAlignment
- // * constructor (VDJAlignmentImmediate).
- // */
- //@SuppressWarnings("unchecked")
- //static AlignmentHit<NucleotideSequence, Allele>[][] toArray(PairedHit[] hits) {
- // AlignmentHit<NucleotideSequence, Allele>[][] hitsArray = new AlignmentHit[hits.length][];
- // for (int i = 0; i < hits.length; ++i)
- // hitsArray[i] = new AlignmentHit[]{hits[i].hit0, hits[i].hit1};
- // return hitsArray;
- //}
-
  /**
  * Calculates normal "sum" score for each hit and sort hits according to this score.
  */

src/main/java/com/milaboratory/mixcr/vdjaligners/VDJCAlignerParameters.java

@@ -190,25 +190,25 @@ public KGeneAlignmentParameters getVJCGeneAlignerParameters(GeneType geneType) {
  }
 
  @JsonProperty("vParameters")
- @JsonSerialize(include = JsonSerialize.Inclusion.NON_NULL)
+ @JsonInclude(JsonInclude.Include.NON_NULL)
  public KGeneAlignmentParameters getVAlignerParameters() {
  return getVJCGeneAlignerParameters(GeneType.Variable);
  }
 
  @JsonProperty("dParameters")
- @JsonSerialize(include = JsonSerialize.Inclusion.NON_NULL)
+ @JsonInclude(JsonInclude.Include.NON_NULL)
  public DAlignerParameters getDAlignerParameters() {
  return (DAlignerParameters) getGeneAlignerParameters(GeneType.Diversity);
  }
 
  @JsonProperty("jParameters")
- @JsonSerialize(include = JsonSerialize.Inclusion.NON_NULL)
+ @JsonInclude(JsonInclude.Include.NON_NULL)
  public KGeneAlignmentParameters getJAlignerParameters() {
  return getVJCGeneAlignerParameters(GeneType.Joining);
  }
 
  @JsonProperty("cParameters")
- @JsonSerialize(include = JsonSerialize.Inclusion.NON_NULL)
+ @JsonInclude(JsonInclude.Include.NON_NULL)
  public KGeneAlignmentParameters getCAlignerParameters() {
  return getVJCGeneAlignerParameters(GeneType.Constant);
  }

src/main/java/com/milaboratory/mixcr/vdjaligners/VDJCAlignerS.java

@@ -338,17 +338,6 @@ public void calculateHits(float minTotalScore, int maxHits) {
  public void alignDC() {
  NucleotideSequence sequence = target.targets[0].getSequence();
 
- if (singleDAligner != null && hasKVAndJHits()) {
- //Alignment of D gene
- int from = vResult.get(0).getAlignment().getSequence2Range().getTo(),
- to = jResult.get(0).getAlignment().getSequence2Range().getFrom();
- List<PreVDJCHit> dResult = from > to ?
- Collections.<PreVDJCHit>emptyList() :
- singleDAligner.align0(sequence, getPossibleDLoci(), from, to);
- dHits = PreVDJCHit.convert(getDGenesToAlign(),
- parameters.getFeatureToAlign(GeneType.Diversity), dResult);
- }
-
  boolean doCAlignment = cAligner != null;
 
  if (parameters.getAllowNoCDR3PartAlignments())
@@ -367,6 +356,21 @@ public void alignDC() {
  cHits = createHits(res.getHits(), parameters.getFeatureToAlign(GeneType.Constant));
  } else
  cHits = new VDJCHit[0];
+
+ if (singleDAligner != null && hasKVAndJHits()) {
+ //Alignment of D gene
+ int from = vResult.get(0).getAlignment().getSequence2Range().getTo(),
+ to = jResult.get(0).getAlignment().getSequence2Range().getFrom();
+ List<PreVDJCHit> dResult = from > to ?
+ Collections.<PreVDJCHit>emptyList() :
+ singleDAligner.align0(sequence, getPossibleDLoci(
+ wrapAlignmentHits(vHits),
+ wrapAlignmentHits(jHits), cHits),
+ from, to);
+ dHits = PreVDJCHit.convert(getDGenesToAlign(),
+ parameters.getFeatureToAlign(GeneType.Diversity), dResult);
+ }
+
  }
 
  public boolean hasNoKVNorKJHits() {
@@ -411,15 +415,6 @@ public float sumScore() {
  return score;
  }
 
- public Chains getPossibleDLoci() {
- Chains chains = new Chains();
- for (AlignmentHit<NucleotideSequence, VDJCGene> vHit : vResult)
- chains = chains.merge(vHit.getRecordPayload().getChains());
- for (AlignmentHit<NucleotideSequence, VDJCGene> jHit : jResult)
- chains = chains.merge(jHit.getRecordPayload().getChains());
- return chains;
- }
-
  public VDJCAlignments toVDJCAlignments(long inputId) {
  EnumMap<GeneType, VDJCHit[]> hits = new EnumMap<>(GeneType.class);