Add support for demuxmix in cell hashing

Author: bbimber

singlecell/api-src/org/labkey/api/singlecell/CellHashingService.java

@@ -24,6 +24,7 @@
 import java.io.IOException;
 import java.util.ArrayList;
 import java.util.Arrays;
+import java.util.Collection;
 import java.util.Collections;
 import java.util.HashSet;
 import java.util.List;
@@ -71,7 +72,7 @@ static public void setInstance(CellHashingService instance)
 
     abstract public boolean usesCiteSeq(SequenceAnalysisJobSupport support, List<SequenceOutputFile> inputFiles) throws PipelineJobException;
 
-    abstract public List<ToolParameterDescriptor> getHashingCallingParams(boolean allowDemuxEm);
+    abstract public List<ToolParameterDescriptor> getHashingCallingParams(boolean allowMethodsNeedingGex);
 
     abstract public Set<String> getHtosForParentReadset(Integer parentReadsetId, File webserverJobDir, SequenceAnalysisJobSupport support, boolean throwIfNotFound) throws PipelineJobException;
 
@@ -316,17 +317,25 @@ public enum CALLING_METHOD
         htodemux(false, false),
         dropletutils(true, true),
         gmm_demux(true, true),
-        demuxem(true, true),
+        demuxem(true, true, true),
+        demuxmix(true, true, true),
         bff_cluster(true, true),
         bff_raw(true, false);
 
         boolean isDefaultRun;
         boolean isDefaultConsensus;
+        boolean requiresH5;
 
         CALLING_METHOD(boolean isDefaultRun, boolean isDefaultConsensus)
+        {
+            this(isDefaultRun, isDefaultConsensus, false);
+        }
+
+        CALLING_METHOD(boolean isDefaultRun, boolean isDefaultConsensus, boolean requiresH5)
         {
             this.isDefaultRun = isDefaultRun;
             this.isDefaultConsensus = isDefaultConsensus;
+            this.requiresH5 = requiresH5;
         }
 
         public boolean isDefaultRun()
@@ -339,6 +348,11 @@ public boolean isDefaultConsensus()
             return isDefaultConsensus;
         }
 
+        public boolean isRequiresH5()
+        {
+            return requiresH5;
+        }
+
         private static List<CALLING_METHOD> getDefaultRunMethods()
         {
             return Arrays.stream(values()).filter(CALLING_METHOD::isDefaultRun).collect(Collectors.toList());
@@ -358,6 +372,22 @@ public static List<String> getDefaultConsensusMethodNames()
         {
             return getDefaultConsensusMethods().stream().map(Enum::name).collect(Collectors.toList());
         }
+
+        public static boolean requiresH5(String methodNames)
+        {
+            methodNames = StringUtils.trimToNull(methodNames);
+            if (methodNames == null)
+            {
+                return false;
+            }
+
+            return requiresH5(Arrays.stream(methodNames.split(";")).map(CALLING_METHOD::valueOf).collect(Collectors.toList()));
+        }
+
+        public static boolean requiresH5(Collection<CALLING_METHOD> methods)
+        {
+            return methods.stream().anyMatch(CALLING_METHOD::isRequiresH5);
+        }
     }
 
     public enum BARCODE_TYPE

singlecell/src/org/labkey/singlecell/CellHashingServiceImpl.java

@@ -95,8 +95,8 @@ public static CellHashingServiceImpl get()
     @Override
     public void prepareHashingForVdjIfNeeded(SequenceOutputHandler.JobContext ctx, final boolean failIfNoHashingReadset) throws PipelineJobException
     {
-        boolean useDemEM = ctx.getParams().optString("methods", "").contains(CellHashingService.CALLING_METHOD.demuxem.name());
-        prepareHashingAndCiteSeqFilesIfNeeded(ctx.getOutputDir(), ctx.getJob(), ctx.getSequenceSupport(), TCR_FIELD, failIfNoHashingReadset, false, true, true, false, useDemEM);
+        boolean needsH5 = CALLING_METHOD.requiresH5(ctx.getParams().optString("methods", ""));
+        prepareHashingAndCiteSeqFilesIfNeeded(ctx.getOutputDir(), ctx.getJob(), ctx.getSequenceSupport(), TCR_FIELD, failIfNoHashingReadset, false, true, true, false, needsH5);
     }
 
     public void prepareHashingAndCiteSeqFilesForFeatureCountsIfNeeded(File sourceDir, PipelineJob job, SequenceAnalysisJobSupport support, String filterField, final boolean failIfNoHashingReadset, final boolean failIfNoCiteSeqReadset) throws PipelineJobException
@@ -436,7 +436,7 @@ private Map<String, Integer> cacheH5Files(PipelineJob job, SequenceAnalysisJobSu
             {
                 if (cachedGenomes.size() > 1)
                 {
-                    throw new PipelineJobException("demuxEM was selected, but more than one cached genome found, cannot infer correct genome. Found: " + StringUtils.join(cachedGenomes, ", "));
+                    throw new PipelineJobException("demuxEM/demuxmix was selected, but more than one cached genome found, cannot infer correct genome. Found: " + StringUtils.join(cachedGenomes, ", "));
                 }
 
                 gexGenomeId = cachedGenomes.iterator().next();
@@ -448,11 +448,11 @@ private Map<String, Integer> cacheH5Files(PipelineJob job, SequenceAnalysisJobSu
                 HashSet<Integer> genomeIds = new HashSet<>(new TableSelector(ti, PageFlowUtil.set("library_id"), filter, new org.labkey.api.data.Sort("-rowid")).getArrayList(Integer.class));
                 if (genomeIds.isEmpty())
                 {
-                    throw new PipelineJobException("demuxEM was selected, but no suitable loupe files were found for GEX readset: " + gexReadset);
+                    throw new PipelineJobException("demuxEM/demuxmix was selected, but no suitable loupe files were found for GEX readset: " + gexReadset);
                 }
                 else if (genomeIds.size() > 1)
                 {
-                    throw new PipelineJobException("demuxEM was selected. Attempting to identify loupe files using GEX readset: " + gexReadset + ", but more than one genome found. Found: " + StringUtils.join(cachedGenomes, ", "));
+                    throw new PipelineJobException("demuxEM/demuxmix was selected. Attempting to identify loupe files using GEX readset: " + gexReadset + ", but more than one genome found. Found: " + StringUtils.join(cachedGenomes, ", "));
                 }
 
                 gexGenomeId = genomeIds.iterator().next();
@@ -479,7 +479,7 @@ else if (genomeIds.size() > 1)
 
                 if (cachedGenomes.size() > 1)
                 {
-                    throw new PipelineJobException("demuxEM was selected, but more than one cached genome found, cannot infer correct genome. Found: " + StringUtils.join(cachedGenomes, ", "));
+                    throw new PipelineJobException("demuxEM/demuxmix was selected, but more than one cached genome found, cannot infer correct genome. Found: " + StringUtils.join(cachedGenomes, ", "));
                 }
 
                 // NOTE: cache this using the source file's genome ID (which might be the TCR library), rather than the GEX genome
@@ -935,7 +935,7 @@ public void processMetrics(SequenceOutputFile so, PipelineJob job, boolean updat
     }
 
     @Override
-    public List<ToolParameterDescriptor> getHashingCallingParams(boolean allowDemuxEm)
+    public List<ToolParameterDescriptor> getHashingCallingParams(boolean allowMethodsNeedingGex)
     {
         List<ToolParameterDescriptor> ret = new ArrayList<>(Arrays.asList(
             ToolParameterDescriptor.create("minCountPerCell", "Min Reads/Cell", null, "ldk-integerfield", null, 5),
@@ -954,7 +954,7 @@ public List<ToolParameterDescriptor> getHashingCallingParams(boolean allowDemuxE
             ToolParameterDescriptor.create("retainRawCountFile", "Retain Raw Counts File", null, "checkbox", null, false)
         ));
 
-        final List<String> allMethods = Arrays.stream(CALLING_METHOD.values()).filter(x -> allowDemuxEm || x != CALLING_METHOD.demuxem).map(Enum::name).toList();
+        final List<String> allMethods = Arrays.stream(CALLING_METHOD.values()).filter(x -> allowMethodsNeedingGex || !x.isRequiresH5()).map(Enum::name).toList();
         ret.add(ToolParameterDescriptor.create("methods", "Calling Methods", "The set of methods to use in calling.", "ldk-simplecombo", new JSONObject()
         {{
             put("multiSelect", true);
@@ -1182,16 +1182,16 @@ public File generateCellHashingCalls(File citeSeqCountOutDir, File outputDir, St
 
         molInfo = ensureLocalCopy(molInfo, outputDir, log, toDelete);
 
-        // h5 file used by demuxEM:
+        // h5 file used by demuxEM/demuxmix:
         File h5 = null;
         if (parameters.h5File != null)
         {
             h5 = ensureLocalCopy(parameters.h5File, outputDir, log, toDelete);
         }
 
-        if (parameters.methods.contains(CALLING_METHOD.demuxem) && h5 == null)
+        if (CALLING_METHOD.requiresH5(parameters.methods) && h5 == null)
         {
-            throw new PipelineJobException("No h5 file provided, but demuxEM was specified");
+            throw new PipelineJobException("No h5 file provided, but demuxEM/demuxmix was specified");
         }
 
         citeSeqCountOutDir = ensureLocalCopy(citeSeqCountOutDir, outputDir, log, toDelete);

singlecell/src/org/labkey/singlecell/analysis/AbstractSingleCellHandler.java

@@ -188,7 +188,7 @@ public void init(JobContext ctx, List<SequenceOutputFile> inputFiles, List<Recor
                 if (step.requiresHashing(ctx))
                 {
                     String methods = step.getProvider().hasParameter("methods") ? step.getProvider().getParameterByName("methods").extractValue(ctx.getJob(), step.getProvider(), step.getStepIdx(), String.class) : null;
-                    if (methods != null && methods.contains(CellHashingService.CALLING_METHOD.demuxem.name()))
+                    if (CellHashingService.CALLING_METHOD.requiresH5(methods))
                     {
                         doH5Caching = true;
                     }

singlecell/src/org/labkey/singlecell/pipeline/singlecell/RunCellHashing.java

@@ -113,7 +113,7 @@ protected Map<Integer, File> prepareCountData(SingleCellOutput output, SequenceO
                 File existingCountMatrixUmiDir = CellHashingService.get().getExistingFeatureBarcodeCountDir(parentReadset, CellHashingService.BARCODE_TYPE.hashing, ctx.getSequenceSupport());
                 params.allowableHtoBarcodes = htosPerReadset;
                 params.keepMarkdown = true;
-                if (params.methods.contains(CellHashingService.CALLING_METHOD.demuxem))
+                if (CellHashingService.CALLING_METHOD.requiresH5(params.methods))
                 {
                     Integer genomeId = wrapper.getSequenceOutputFile() == null ? null : wrapper.getSequenceOutputFile().getLibrary_id();
                     if (genomeId == null)