Merge branch 'release/v2.1.3'

CHANGELOG

@@ -1,4 +1,15 @@
 
+MiXCR 2.1.3 ( 7 Jun 2017)
+========================
+
+-- Minimal thresholds added for V and J alignments in `extendAlignments`
+-- `minimalNOverlap` in `assemblePartial` changed to `7`
+-- Information on k-mer diversity added to alignPartial report file
+-- Fix: MiXCR now correctly detects `*.fa` files as FASTA formatted
+-- Detainls of V/J Extension during extendAlignments are now added to target description, and can 
+   be retrieved using `-targetDescriptions` field in `exportAlignments`
+
+
 MiXCR 2.1.2 (20 Apr 2017)
 ========================
 

pom.xml

@@ -32,7 +32,7 @@
 
     <groupId>com.milaboratory</groupId>
     <artifactId>mixcr</artifactId>
-    <version>2.1.2</version>
+    <version>2.1.3</version>
     <packaging>jar</packaging>
     <name>MiXCR</name>
 
@@ -44,14 +44,14 @@
 
     <properties>
         <project.build.sourceEncoding>UTF-8</project.build.sourceEncoding>
-        <milib.version>1.7.2</milib.version>
+        <milib.version>1.8</milib.version>
     </properties>
 
     <dependencies>
         <dependency>
             <groupId>io.repseq</groupId>
             <artifactId>repseqio</artifactId>
-            <version>1.2.7</version>
+            <version>1.2.8</version>
         </dependency>
 
         <dependency>
@@ -102,6 +102,32 @@
 
     <build>
         <plugins>
+            <plugin>
+                <!-- Workaround maven not being able to set a property conditionally -->
+                <groupId>org.apache.maven.plugins</groupId>
+                <artifactId>maven-antrun-plugin</artifactId>
+                <version>1.8</version>
+                <executions>
+                    <execution>
+                        <phase>validate</phase>
+                        <goals>
+                            <goal>run</goal>
+                        </goals>
+                        <configuration>
+                            <exportAntProperties>true</exportAntProperties>
+                            <target>
+                                <exec executable="hostname" failonerror="false" outputproperty="hostname"/>
+                                <condition property="hostname" value="unknown">
+                                    <not>
+                                        <isset property="hostname"/>
+                                    </not>
+                                </condition>
+                            </target>
+                        </configuration>
+                    </execution>
+                </executions>
+            </plugin>
+
             <plugin>
                 <groupId>org.codehaus.mojo</groupId>
                 <artifactId>buildnumber-maven-plugin</artifactId>
@@ -123,6 +149,7 @@
                         <configuration>
                             <properties>
                                 <branch>${scmBranch}</branch>
+                                <host>${hostname}</host>
                             </properties>
                         </configuration>
                     </execution>

src/main/java/com/milaboratory/mixcr/cli/ActionAlign.java

@@ -394,7 +394,7 @@ public boolean printWarnings() {
         }
 
         public Chains getChains() {
-            return Util.parseLoci(chains);
+            return Chains.parse(chains);
         }
 
         public boolean getWriteAllResults() {
@@ -414,7 +414,7 @@ public SequenceReaderCloseable<? extends SequenceRead> createReader() throws IOE
                 return new PairedFastqReader(parameters.get(0), parameters.get(1), true);
             else {
                 String[] s = parameters.get(0).split("\\.");
-                if (s[s.length - 1].equals("fasta"))
+                if (s[s.length - 1].equals("fasta") || s[s.length - 1].equals("fa"))
                     return new FastaSequenceReaderWrapper(
                             new FastaReader<>(parameters.get(0), NucleotideSequence.ALPHABET),
                             true

src/main/java/com/milaboratory/mixcr/cli/ActionExportAlignmentsPretty.java

@@ -299,7 +299,7 @@ TLongHashSet getReadIds() {
         }
 
         public Chains getChain() {
-            return Util.parseLoci(chain);
+            return Chains.parse(chain);
         }
 
         @SuppressWarnings("unchecked")

src/main/java/com/milaboratory/mixcr/cli/ActionExportClonesPretty.java

@@ -150,7 +150,7 @@ public static final class AParameters extends ActionParameters {
         public String csContain = null;
 
         public Chains getChain() {
-            return Util.parseLoci(chain);
+            return Chains.parse(chain);
         }
 
         public Filter<Clone> getFilter() {

src/main/java/com/milaboratory/mixcr/cli/ActionExportParameters.java

@@ -102,9 +102,9 @@ public Chains getChains() {
             if (!chains.equals("ALL"))
                 throw new ParameterException("Use -c without -l parameter.");
             System.out.println("WARNING: using of -l (--loci) option is deprecated; use -c (--chains) instead.");
-            return Util.parseLoci(chains_legacy);
+            return Chains.parse(chains_legacy);
         }
-        return Util.parseLoci(chains);
+        return Chains.parse(chains);
     }
 
     @SuppressWarnings("unchecked")

src/main/java/com/milaboratory/mixcr/cli/ActionExtendAlignments.java

@@ -37,6 +37,7 @@ public void go(ActionHelper helper) throws Exception {
             AlignmentExtender extender = new AlignmentExtender(parameters.getChains(), parameters.extensionQuality,
                     reader.getParameters().getVAlignerParameters().getParameters().getScoring(),
                     reader.getParameters().getJAlignerParameters().getParameters().getScoring(),
+                    parameters.minimalVScore, parameters.minimalJScore,
                     ReferencePoint.parse(parameters.vAnchorPoint),
                     ReferencePoint.parse(parameters.jAnchorPoint));
 
@@ -89,14 +90,22 @@ private static final class ExtendCD3Parameters extends ActionParametersWithOutpu
                 names = {"-q", "--quality"})
         public byte extensionQuality = 30;
 
-        @Parameter(description = "V extension anchor point",
+        @Parameter(description = "V extension anchor point.",
                 names = {"--v-anchor"})
         public String vAnchorPoint = "CDR3Begin";
 
-        @Parameter(description = "J extension anchor point",
+        @Parameter(description = "J extension anchor point.",
                 names = {"--j-anchor"})
         public String jAnchorPoint = "CDR3End";
 
+        @Parameter(description = "Minimal score of V alignment to perform left extension.",
+                names = {"--min-v-score"})
+        public int minimalVScore = 100;
+
+        @Parameter(description = "Minimal score of J alignment to perform right extension.",
+                names = {"--min-j-score"})
+        public int minimalJScore = 70;
+
         private String getInput() {
             return parameters.get(0);
         }
@@ -106,7 +115,7 @@ private String getOutput() {
         }
 
         public Chains getChains() {
-            return Util.parseLoci(chains);
+            return Chains.parse(chains);
         }
 
         @Override

src/main/java/com/milaboratory/mixcr/cli/ActionFilterAlignments.java

@@ -158,7 +158,7 @@ protected List<String> getOutputFiles() {
         }
 
         public Chains getChains() {
-            return Util.parseLoci(chains);
+            return Chains.parse(chains);
         }
 
         public VDJCAlignmentsReader getInput() throws IOException {

src/main/java/com/milaboratory/mixcr/cli/ActionSortAlignments.java

@@ -15,6 +15,7 @@
 import com.milaboratory.mixcr.basictypes.VDJCAlignmentsReader;
 import com.milaboratory.mixcr.basictypes.VDJCAlignmentsWriter;
 import com.milaboratory.mixcr.vdjaligners.VDJCAlignerParameters;
+import com.milaboratory.util.ObjectSerializer;
 import com.milaboratory.util.SmartProgressReporter;
 import com.milaboratory.util.Sorter;
 import com.milaboratory.util.TempFileManager;
@@ -70,7 +71,7 @@ public int compare(VDJCAlignments o1, VDJCAlignments o2) {
         }
     };
 
-    private static final class VDJCAlignmnetsSerializer implements Sorter.ObjectSerializer<VDJCAlignments> {
+    private static final class VDJCAlignmnetsSerializer implements ObjectSerializer<VDJCAlignments> {
         final VDJCAlignerParameters parameters;
         final List<VDJCGene> usedAlleles;
 

src/main/java/com/milaboratory/mixcr/cli/Main.java

@@ -97,23 +97,19 @@ public static void main(String... args) throws Exception {
                 new ActionSortAlignments());
 
         // Adding version info callback
-        main.setVersionInfoCallback(new Runnable() {
-            @Override
-            public void run() {
-                System.err.print(
-                        MiXCRVersionInfo.get().getVersionString(
-                                MiXCRVersionInfo.OutputType.ToConsole));
-                System.err.println();
-                System.err.println("Library search path:");
-                for (VDJCLibraryRegistry.LibraryResolver resolvers : VDJCLibraryRegistry.getDefault()
-                        .getLibraryResolvers()) {
-                    if (resolvers instanceof VDJCLibraryRegistry.ClasspathLibraryResolver)
-                        System.out.println("- built-in libraries");
-                    if (resolvers instanceof VDJCLibraryRegistry.FolderLibraryResolver)
-                        System.out.println("- " + ((VDJCLibraryRegistry.FolderLibraryResolver) resolvers).getPath());
-                }
-            }
-        });
+        main.setVersionInfoCallback(
+                new Runnable() {
+                    @Override
+                    public void run() {
+                        printVersion(false);
+                    }
+                },
+                new Runnable() {
+                    @Override
+                    public void run() {
+                        printVersion(true);
+                    }
+                });
 
         // Executing main method
         JCommanderBasedMain.ProcessResult processResult = main.main(args);
@@ -122,4 +118,19 @@ public void run() {
         if (processResult == JCommanderBasedMain.ProcessResult.Error)
             System.exit(1);
     }
+
+    static void printVersion(boolean full) {
+        System.err.print(
+                MiXCRVersionInfo.get().getVersionString(
+                        MiXCRVersionInfo.OutputType.ToConsole, full));
+        System.err.println();
+        System.err.println("Library search path:");
+        for (VDJCLibraryRegistry.LibraryResolver resolvers : VDJCLibraryRegistry.getDefault()
+                .getLibraryResolvers()) {
+            if (resolvers instanceof VDJCLibraryRegistry.ClasspathLibraryResolver)
+                System.out.println("- built-in libraries");
+            if (resolvers instanceof VDJCLibraryRegistry.FolderLibraryResolver)
+                System.out.println("- " + ((VDJCLibraryRegistry.FolderLibraryResolver) resolvers).getPath());
+        }
+    }
 }

src/main/java/com/milaboratory/mixcr/cli/Util.java

@@ -48,27 +48,7 @@ private Util() {
     }
 
     public static final DecimalFormat PERCENT_FORMAT = new DecimalFormat("#.##");
-
-    public static Chains parseLoci(String lociString) {
-        String[] split = lociString.split(",");
-        Chains chains = new Chains();
-        for (String s : split)
-            chains = chains.merge(parseLocus(s));
-        return chains;
-    }
-
-    private static Chains parseLocus(String value) {
-        switch (value.toLowerCase().trim()) {
-            case "tcr":
-                return Chains.TCR;
-            case "ig":
-                return Chains.IG;
-            case "all":
-                return Chains.ALL;
-        }
-        return new Chains(value);
-    }
-
+
     public static void writeReport(String input, String output,
                                    String commandLineArguments,
                                    String reportFileName,

src/main/java/com/milaboratory/mixcr/export/FieldExtractors.java

@@ -489,6 +489,26 @@ protected String extract(VDJCAlignments object) {
                 }
             });
 
+            descriptorsList.add(new PL_A("-targetDescriptions", "Export target descriptions", "Target descriptions", "targetDescriptions") {
+                @Override
+                protected String extract(VDJCAlignments object) {
+                    String[] ds = object.getTargetDescriptions();
+                    if (ds == null || ds.length == 0) {
+                        char[] commas = new char[object.numberOfTargets() - 1];
+                        Arrays.fill(commas, ',');
+                        return new String(commas);
+                    }
+                    StringBuilder sb = new StringBuilder();
+                    for (int i = 0; ; i++) {
+                        sb.append(ds[i]);
+                        if (i == ds.length - 1)
+                            break;
+                        sb.append(',');
+                    }
+                    return sb.toString();
+                }
+            });
+
             descriptorsList.add(alignmentsToClone("-cloneId", "To which clone alignment was attached.", false));
             descriptorsList.add(alignmentsToClone("-cloneIdWithMappingType", "To which clone alignment was attached with additional info on mapping type.", true));
             descriptorsList.add(new AbstractField<Clone>(Clone.class, "-readIds", "Read IDs aggregated by clone.") {

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

@@ -63,6 +63,7 @@ public class PartialAlignmentsAssembler implements AutoCloseable, ReportWriter {
             rightParts = new AtomicLong(),
             noKMer = new AtomicLong(),
             wildCardsInKMer = new AtomicLong(),
+            kMerDiversity = new AtomicLong(),
             total = new AtomicLong(),
             overlapped = new AtomicLong(),
             totalWritten = new AtomicLong(),
@@ -468,6 +469,7 @@ public void writeReport(ReportHelper helper) {
         helper.writePercentAndAbsoluteField("Alignments already with CDR3 (no overlapping is performed)", containsCDR3, total);
         helper.writePercentAndAbsoluteField("Successfully overlapped alignments", overlapped, total);
         helper.writePercentAndAbsoluteField("Left parts with too small N-region (failed to extract k-mer)", noKMer, total);
+        helper.writeField("Extracted k-mer diversity", kMerDiversity);
         helper.writePercentAndAbsoluteField("Dropped due to wildcard in k-mer", wildCardsInKMer, total);
         helper.writePercentAndAbsoluteField("Dropped due to too short NRegion parts in overlap", droppedSmallOverlapNRegion, total);
         helper.writePercentAndAbsoluteField("Dropped overlaps with empty N region due to no complete NDN coverage", droppedNoNRegion, total);
@@ -553,8 +555,10 @@ private boolean addLeftToIndex(VDJCAlignments alignment) {
             }
 
             List<KMerInfo> ids = kToIndexLeft.get(kmer);
-            if (ids == null)
+            if (ids == null) {
                 kToIndexLeft.put(kmer, ids = new ArrayList<>(1));
+                kMerDiversity.incrementAndGet();
+            }
             ids.add(new KMerInfo(alignment, kFrom, leftTargetId));
         }
 
@@ -568,7 +572,7 @@ private static long kMer(NucleotideSequence seq, int from, int length) {
             byte c = seq.codeAt(j);
             if (NucleotideSequence.ALPHABET.isWildcard(c))
                 return -1;
-            kmer = (kmer << 2|c);
+            kmer = (kmer << 2 | c);
         }
         return kmer;
     }