MANTA-1247 Add a configurable option graphNodeMaxEdgeCount

CHANGELOG.md

@@ -1,5 +1,7 @@
 ## Unreleased
 ### Added
+- Add a configurable option `graphNodeMaxEdgeCount` (MANTA-1247)
+  - During SV candidate generation, if both nodes of an edge have an edge count higher than the configured value, that edge will be skipped for evaluation. 
 
 ### Changed
 - Test for unsupported BAM SEQ format (MANTA-1265)

src/c++/lib/applications/GenerateSVCandidates/EdgeOptionsParser.cpp

@@ -45,7 +45,9 @@ getOptionsDescription(
     ("bin-count", po::value(&opt.binCount)->default_value(opt.binCount),
      "Specify how many bins the SV candidate problem should be divided into, where bin-index can be used to specify which bin to solve")
     ("bin-index", po::value(&opt.binIndex)->default_value(opt.binIndex),
-     "specify which bin to solve when the SV candidate problem is subdivided into bins. Value must bin in [0,bin-count)")
+     "Specify which bin to solve when the SV candidate problem is subdivided into bins. Value must bin in [0,bin-count)")
+    ("max-edge-count", po::value(&opt.graphNodeMaxEdgeCount)->default_value(opt.graphNodeMaxEdgeCount),
+     "Specify the maximum number of edge count. If both nodes of an edge have an edge count higher than this, the edge is skipped for evaluation.")
     (locusIndexKey, po::value<std::string>(),
      "Instead of solving for all SV candidates in a bin, solve for candidates of a particular locus or edge."
      " If this argument is specified then bin-index is ignored."

src/python/bin/configManta.py.ini

@@ -15,6 +15,10 @@ rnaMinCandidateVariantSize = 1000
 # Note that one supporting read pair or split read usually equals one observation, but evidence is sometimes downweighted.
 minEdgeObservations = 3
 
+# If both nodes of an edge have an edge count higher than this, then skip evaluation of the edge.
+# Set to 0 to turn this filtration off
+graphNodeMaxEdgeCount = 10
+
 # Run discovery and candidate reporting for all SVs/indels with at least this
 # many spanning support observations
 minCandidateSpanningCount = 3

src/python/lib/mantaWorkflow.py

@@ -541,6 +541,7 @@ def runHyGen(self, taskPrefix="", dependencies=None) :
         hygenCmd.extend(["--graph-file",graphPath])
         hygenCmd.extend(["--bin-index", str(binId)])
         hygenCmd.extend(["--bin-count", str(self.params.nonlocalWorkBins)])
+        hygenCmd.extend(["--max-edge-count", str(self.params.graphNodeMaxEdgeCount)])
         hygenCmd.extend(["--min-candidate-sv-size", self.params.minCandidateVariantSize])
         hygenCmd.extend(["--min-candidate-spanning-count", self.params.minCandidateSpanningCount])
         hygenCmd.extend(["--min-scored-sv-size", self.params.minScoredVariantSize])