public void GetAlignmentSet()
{
var reads = new List <Read>();
reads.Add(CreateRead("chr1", "ACGT", 10, "read1", matePosition: 10));
reads.Add(CreateRead("chr1", "ACGT", 10, "read2", matePosition: 10));
reads.Add(CreateRead("chr1", "ACGT", 10, "read1", read2: true, matePosition: 10)); // mate
reads.Add(CreateRead("chr1", "ACGT", 10, "read_notmapped", isMapped: false, isProperPair: false, matePosition: 10));
reads.Add(CreateRead("chr1", "ACGT", 10, "read3", isProperPair: false, read2: true, matePosition: 10)); // mate
reads.Add(CreateRead("chr1", "ACGT", 10, "read2", read2: true, matePosition: 10));
var extractor = new MockAlignmentExtractor(reads);
var config = new AlignmentSourceConfig {
MinimumMapQuality = 10, OnlyUseProperPairs = false
};
var source = new AlignmentSource(extractor, null, config);
Read read;
var numSets = 0;
while ((read = source.GetNextRead()) != null)
{
numSets++;
Assert.False(read.Name == "read_notmapped");
}
Assert.Equal(numSets, 5);
}
private void ExecuteGroupingTest(List <Read> reads, List <int> expectedGroupMemberships, IEnumerable <Tuple <int, string, string> > variants)
{
var variantSites = new List <VariantSite>();
foreach (var variant in variants)
{
variantSites.Add(new VariantSite(variant.Item1)
{
VcfReferenceAllele = variant.Item2, VcfAlternateAllele = variant.Item3
});
}
var alignmentExtractor = new MockAlignmentExtractor(reads);
var veadSource = new VeadGroupSource(alignmentExtractor, new BamFilterParameters()
{
MinimumMapQuality = 20
}, false, "");
var vcfNeighborhood = new VcfNeighborhood(new VariantCallingParameters(), "chr1", new VariantSite(120), new VariantSite(121), "T")
{
VcfVariantSites = variantSites
};
vcfNeighborhood.SetRangeOfInterest();
var veadGroups = veadSource.GetVeadGroups(vcfNeighborhood).ToList();
Assert.Equal(expectedGroupMemberships.Count, veadGroups.Count());
for (var i = 0; i < veadGroups.Count(); i++)
{
Assert.Equal(expectedGroupMemberships[i], veadGroups[i].NumVeads);
}
}
private void ExecuteTest_ProcessReadsSeparately(bool doMate, bool doStitch)
{
var read = CreateRead("chr1", "ACGT", 10, "read1", matePosition: 100);
var readMate = CreateRead("chr1", "ACGT", 100, "read1", read2: true, matePosition: 10);
var extractor = new MockAlignmentExtractor(new List <Read>()
{
read
});
var mateFinder = new Mock <IAlignmentMateFinder>();
mateFinder.Setup(f => f.GetMate(It.IsAny <Read>())).Returns(readMate);
var stitcher = new Mock <IAlignmentStitcher>();
var config = new AlignmentSourceConfig {
MinimumMapQuality = 10, OnlyUseProperPairs = false
};
var source = new AlignmentSource(extractor, doMate ? mateFinder.Object : null, config);
var alignmentSet = source.GetNextRead();
// TODO what do we really want to test here now that we're not pairing?
}
public void SupplementSupportWithClippedReads()
{
// In this test we create reads that are either normal or clipped (identified by "clip_" in their name)
// This test does not take cigar data into account.
var mockClippedReadComparator = new Mock <IMNVClippedReadComparator>();
// Mock read comparator returns true if read name starts with c
mockClippedReadComparator.Setup(x => x.DoesClippedReadSupportMNV(It.IsAny <Read>(), It.IsAny <CalledAllele>()))
.Returns((Read read, CalledAllele allele) => read.Name[0] == 'c' ? true : false);
var reads = new List <Read>();
reads.Add(CreateRead("chr1", "ACGT", 3, "read4"));
reads.Add(CreateRead("chr1", "ACGT", 3, "clip_read4", matePosition: 3)); // +1 not in neighborhood, but still gets counted because mocked ClippedReadComparator
reads.Add(CreateRead("chr1", "ACGT", 12, "read1", matePosition: 10));
reads.Add(CreateRead("chr1", "ACGT", 12, "read2", matePosition: 10));
reads.Add(CreateRead("chr1", "ACGT", 12, "read1", read2: true, matePosition: 10));
reads.Add(CreateRead("chr1", "ACGT", 12, "read_notmapped", isMapped: false, isProperPair: false, matePosition: 10));
reads.Add(CreateRead("chr1", "ACGT", 12, "read3", isProperPair: false, read2: true, matePosition: 10));
reads.Add(CreateRead("chr1", "ACGT", 12, "read2", read2: true, matePosition: 10));
reads.Add(CreateRead("chr1", "ACGT", 12, "clip_read1", matePosition: 10)); // +1 clipped read
reads.Add(CreateRead("chr1", "ACGT", 12, "clip_read2", matePosition: 10)); // +1 clipped read
reads.Add(CreateRead("chr1", "ACGT", 12, "clip_read1", read2: true, matePosition: 10)); // +1 clipped read
reads.Add(CreateRead("chr1", "ACGT", 12, "clip_read_notmapped", isMapped: false, isProperPair: false, matePosition: 10)); // +1 clipped read
reads.Add(CreateRead("chr1", "ACGT", 12, "clip_read3", isProperPair: false, read2: true, matePosition: 10)); // +1 clipped read
reads.Add(CreateRead("chr1", "ACGT", 12, "clip_read2", read2: true, matePosition: 10)); // +1 clipped read
reads.Add(CreateRead("chr1", "ACGT", 30, "read5"));
reads.Add(CreateRead("chr1", "ACGT", 30, "clip_read5", matePosition: 30)); // not in neighborhood, not counted
var mockAlignmentExtractor = new MockAlignmentExtractor(reads);
int qNoiseLevel = 20;
int maxQscore = 100;
int minMNVsize = 6;
MNVSoftClipSupportFinder mnvClippedSupportFinder = new MNVSoftClipSupportFinder(mockAlignmentExtractor, mockClippedReadComparator.Object, qNoiseLevel, maxQscore, minMNVsize);
var mnv1 = TestHelper.CreateDummyAllele("chr1", 10, "AAAAAA", "CCC", 2000, 50);
var neighbor1 = new VcfNeighborhood(0, "chr", new VariantSite(10000), new VariantSite(200000))
{
VcfVariantSites = new List <VariantSite>
{
new VariantSite(10)
{
VcfReferenceAllele = "A", VcfAlternateAllele = "C", ReferenceName = "chr"
},
new VariantSite(25)
{
VcfReferenceAllele = "T", VcfAlternateAllele = "G", ReferenceName = "chr"
},
},
};
var callableNbhd = new CallableNeighborhood(neighbor1, new VariantCallingParameters(), null);
callableNbhd.AddAcceptedPhasedVariant(mnv1);
Assert.Equal(50, callableNbhd.CandidateVariants[0].AlleleSupport);
mnvClippedSupportFinder.SupplementSupportWithClippedReads(callableNbhd);
Assert.Equal(57, callableNbhd.CandidateVariants[0].AlleleSupport);
}
public AmpliconTestFactory(string referenceSequence, bool sourceIsStitched = false)
{
_referenceSequence = referenceSequence;
ChrInfo = new ChrReference()
{
Name = "chr7",
Sequence = _referenceSequence
};
AlignmentExtractor = new MockAlignmentExtractor(ChrInfo, sourceIsStitched);
}
public AmpliconTestFactory(string referenceSequence, bool stitchReads)
{
_referenceSequence = referenceSequence;
ChrInfo = new ChrReference()
{
Name = "chr7",
Sequence = _referenceSequence
};
_MAE = new MockAlignmentExtractor(ChrInfo, stitchReads);
}
public void GetAlignmentSet()
{
var reads = new List <Read>();
reads.Add(CreateRead("chr1", "ACGT", 1, "read1"));
reads.Add(CreateRead("chr1", "ACGT", 1, "read2"));
reads.Add(CreateRead("chr1", "ACGT", 1, "read1_mate"));
reads.Add(CreateRead("chr1", "ACGT", 1, "read3_unmapped", isMapped: false, isProperPair: false));
reads.Add(CreateRead("chr1", "ACGT", 1, "read3_mate", isProperPair: false));
reads.Add(CreateRead("chr1", "ACGT", 1, "read2_mate"));
var extractor = new MockAlignmentExtractor(reads);
var mateFinder = new Mock <IAlignmentMateFinder>();
mateFinder.Setup(f => f.GetMate(It.IsAny <Read>())).Returns(
(Read r) =>
r.Name.EndsWith("_mate")
? reads.FirstOrDefault(o => o.Name == r.Name.Replace("_mate", string.Empty))
: null);
var stitcher = new Mock <IAlignmentStitcher>();
stitcher.Setup(s => s.TryStitch(It.IsAny <AlignmentSet>())).Callback((AlignmentSet s) => s.IsStitched = true);
var config = new AlignmentSourceConfig {
MinimumMapQuality = 10, OnlyUseProperPairs = false
};
var source = new AlignmentSource(extractor, mateFinder.Object, stitcher.Object, config);
AlignmentSet set;
var numSets = 0;
while ((set = source.GetNextAlignmentSet()) != null)
{
numSets++;
if (!set.IsFullPair)
{
Assert.True(set.PartnerRead1.Name.Equals("read3_mate"));
Assert.True(set.PartnerRead2 == null);
Assert.False(set.IsStitched);
}
else
{
Assert.True(set.PartnerRead2.Name.Equals(set.PartnerRead1.Name + "_mate"));
Assert.True(set.IsStitched);
}
}
Assert.Equal(numSets, 3);
}
private void ExecuteTest_ProcessReadsSeparately(bool doMate, bool doStitch)
{
var read = CreateRead("chr1", "ACGT", 1, "read1");
var readMate = CreateRead("chr1", "ACGT", 10, "read1mate");
var extractor = new MockAlignmentExtractor(new List <Read>()
{
read
});
var mateFinder = new Mock <IAlignmentMateFinder>();
mateFinder.Setup(f => f.GetMate(It.IsAny <Read>())).Returns(readMate);
var stitcher = new Mock <IAlignmentStitcher>();
var config = new AlignmentSourceConfig {
MinimumMapQuality = 10, OnlyUseProperPairs = false
};
var source = new AlignmentSource(extractor, doMate ? mateFinder.Object : null, doStitch ? stitcher.Object : null, config);
var alignmentSet = source.GetNextAlignmentSet();
// verify stitching relies on pairing
stitcher.Verify(s => s.TryStitch(It.IsAny <AlignmentSet>()), Times.Never);
if (!doMate)
{
TestHelper.CompareReads(read, alignmentSet.PartnerRead1);
Assert.Equal(null, alignmentSet.PartnerRead2);
Assert.Equal(1, alignmentSet.ReadsForProcessing.Count);
TestHelper.CompareReads(read, alignmentSet.ReadsForProcessing.First());
}
else
{
TestHelper.CompareReads(read, alignmentSet.PartnerRead1);
TestHelper.CompareReads(readMate, alignmentSet.PartnerRead2);
Assert.Equal(2, alignmentSet.ReadsForProcessing.Count);
TestHelper.CompareReads(read, alignmentSet.ReadsForProcessing.First());
TestHelper.CompareReads(readMate, alignmentSet.ReadsForProcessing.Last());
}
}
public void GetVeads()
{
var vcfNeighborhood = new VcfNeighborhood(new VariantCallingParameters(), "chr1", new VariantSite(10000), new VariantSite(200000), "T")
{
VcfVariantSites = new List <VariantSite>
{
new VariantSite(100)
{
VcfReferenceAllele = "A", VcfAlternateAllele = "C"
},
new VariantSite(400)
{
VcfReferenceAllele = "A", VcfAlternateAllele = "C"
},
new VariantSite(505)
{
VcfReferenceAllele = "A", VcfAlternateAllele = "C"
},
new VariantSite(703)
{
VcfReferenceAllele = "A", VcfAlternateAllele = "T"
},
new VariantSite(800)
{
VcfReferenceAllele = "A", VcfAlternateAllele = "C"
},
}
};
var reads = new List <Read>();
reads.Add(CreateRead("chr1", "ACGT", 10)); // Before neighborhood
reads.Add(CreateRead("chr1", "ACGT", 96)); // Ends right before neighborhood's first variant site
reads.Add(CreateRead("chr1", "ACGT", 100)); // Match (100)
reads.Add(CreateRead("chr1", "ACGT", 300)); // Within neighborhood but no VariantSite
reads.Add(CreateRead("chr1", "ACGT", 400, qualityForAll: 19)); // Within neighbhorhood but low quals
reads.Add(CreateRead("chr1", "ACGT", 500)); // Within neighborhood but no VariantSite (ends right before 505)
reads.Add(CreateRead("chr1", "ACGT", 700)); // Match (703)
reads.Add(CreateRead("chr1", "ACGT", 800)); // Match (800)
reads.Add(CreateRead("chr1", "ACGT", 805)); // Past neighborhood
reads.Add(CreateRead("chr1", "ACGT", 900)); // Past neighborhood
reads.Add(CreateRead("chr2", "ACGT", 100)); // Wrong chromosome
vcfNeighborhood.SetRangeOfInterest();
var alignmentExtractor = new MockAlignmentExtractor(reads);
var veadSource = new VeadGroupSource(alignmentExtractor,
new BamFilterParameters()
{
MinimumMapQuality = 20
}, false, "");
var veadGroups = veadSource.GetVeadGroups(vcfNeighborhood);
// Collect all reads that could relate to the neighborhood
// - Skip anything that has quality less than MinimumMapQuality
// - Skip anything that ends before neighborhood begins
// - Stop collecting once we've passed the end of the neighborhood
// We should have collected the reads at 100, 700, and 800.
Assert.Equal(801, vcfNeighborhood.LastPositionOfInterestWithLookAhead);
Assert.Equal(3, veadGroups.Count());
Assert.Equal(1, veadGroups.Count(v => v.RepresentativeVead.Name.EndsWith("100")));
Assert.Equal(1, veadGroups.Count(v => v.RepresentativeVead.Name.EndsWith("700")));
Assert.Equal(1, veadGroups.Count(v => v.RepresentativeVead.Name.EndsWith("800")));
Assert.Equal(0, veadGroups.Count(v => v.RepresentativeVead.Name.EndsWith("805")));
Assert.Equal(0, veadGroups.Count(v => v.RepresentativeVead.Name.EndsWith("900")));
foreach (var veadGroup in veadGroups)
{
Assert.Equal(1, veadGroup.NumVeads);
}
vcfNeighborhood.VcfVariantSites.Add(
new VariantSite(790)
{
VcfReferenceAllele = "ACAGTGAAAGACTTGTGAC", VcfAlternateAllele = "C"
});
vcfNeighborhood.SetRangeOfInterest();
Assert.Equal(809, vcfNeighborhood.LastPositionOfInterestWithLookAhead);
alignmentExtractor = new MockAlignmentExtractor(reads);
veadSource = new VeadGroupSource(alignmentExtractor,
new BamFilterParameters()
{
MinimumMapQuality = 20
}, false, "");
veadGroups = veadSource.GetVeadGroups(vcfNeighborhood);
Assert.Equal(3, veadGroups.Count());
Assert.Equal(1, veadGroups.Count(v => v.RepresentativeVead.Name.EndsWith("100")));
Assert.Equal(1, veadGroups.Count(v => v.RepresentativeVead.Name.EndsWith("700")));
Assert.Equal(1, veadGroups.Count(v => v.RepresentativeVead.Name.EndsWith("800")));
Assert.Equal(0, veadGroups.Count(v => v.RepresentativeVead.Name.EndsWith("805")));
Assert.Equal(0, veadGroups.Count(v => v.RepresentativeVead.Name.EndsWith("900")));
// Boundary case - read ends exactly at neighborhood's first variant site
reads = new List <Read>();
reads.Add(CreateRead("chr1", "ACGT", 10)); // Before neighborhood
reads.Add(CreateRead("chr1", "ACGT", 96)); // Ends right before neighborhood's first variant site
reads.Add(CreateRead("chr1", "ACGT", 97)); // Ends exactly at neighborhood's first variant site
alignmentExtractor = new MockAlignmentExtractor(reads);
veadSource = new VeadGroupSource(alignmentExtractor, new BamFilterParameters()
{
MinimumMapQuality = 20
}, false, "");
veadGroups = veadSource.GetVeadGroups(vcfNeighborhood);
// The veadgroup for 97 should be the only one
Assert.Equal(1, veadGroups.Count());
Assert.Equal(1, veadGroups.Count(v => v.RepresentativeVead.Name.EndsWith("97")));
foreach (var veadGroup in veadGroups)
{
Assert.Equal(1, veadGroup.NumVeads);
}
}
public static ISomaticVariantCaller CreateMockVariantCaller(VcfFileWriter vcfWriter, ApplicationOptions options, ChrReference chrRef, MockAlignmentExtractor mockAlignmentExtractor, IStrandBiasFileWriter biasFileWriter = null, string intervalFilePath = null)
{
var config = new AlignmentSourceConfig
{
MinimumMapQuality = options.MinimumMapQuality,
OnlyUseProperPairs = options.OnlyUseProperPairs,
};
//var mateFinder = options.StitchReads ? new AlignmentMateFinder() : null;
AlignmentMateFinder mateFinder = null;
var alignmentSource = new AlignmentSource(mockAlignmentExtractor, mateFinder, config);
var variantFinder = new CandidateVariantFinder(options.MinimumBaseCallQuality, options.MaxSizeMNV, options.MaxGapBetweenMNV, options.CallMNVs);
var coverageCalculator = new CoverageCalculator();
var alleleCaller = new AlleleCaller(new VariantCallerConfig
{
IncludeReferenceCalls = options.OutputgVCFFiles,
MinVariantQscore = options.MinimumVariantQScore,
MaxVariantQscore = options.MaximumVariantQScore,
VariantQscoreFilterThreshold = options.FilteredVariantQScore > options.MinimumVariantQScore ? options.FilteredVariantQScore : (int?)null,
MinCoverage = options.MinimumDepth,
MinFrequency = options.MinimumFrequency,
EstimatedBaseCallQuality = options.AppliedNoiseLevel == -1 ? options.MinimumBaseCallQuality : options.AppliedNoiseLevel,
StrandBiasModel = options.StrandBiasModel,
StrandBiasFilterThreshold = options.StrandBiasAcceptanceCriteria,
FilterSingleStrandVariants = options.FilterOutVariantsPresentOnlyOneStrand,
ChrReference = chrRef
},
coverageCalculator: coverageCalculator,
variantCollapser: options.Collapse ? new VariantCollapser(null, coverageCalculator) : null);
var stateManager = new RegionStateManager(
expectStitchedReads: mockAlignmentExtractor.SourceIsStitched,
trackOpenEnded: options.Collapse, trackReadSummaries: options.CoverageMethod == CoverageMethod.Approximate);
//statmanager is an allele source
Assert.Equal(0, stateManager.GetAlleleCount(1, AlleleType.A, DirectionType.Forward));
return(new SomaticVariantCaller(
alignmentSource,
variantFinder,
alleleCaller,
vcfWriter,
stateManager,
chrRef,
null,
biasFileWriter));
}
public void Filtering()
{
//bool isMapped = true, bool isPrimaryAlignment = true, bool isProperPair = true, bool isDuplicate = false
var reads = new List <Read>();
reads.Add(CreateRead("chr1", "AAA", 1, "filtered", false, true, true, false, 10));
reads.Add(CreateRead("chr1", "AAA", 1, "filtered", true, false, true, false, 10));
reads.Add(CreateRead("chr1", "AAA", 1, "filtered_properpair", true, true, false, false, 10));
reads.Add(CreateRead("chr1", "AAA", 1, "filtered", true, true, true, true, 10));
reads.Add(CreateRead("chr1", "AAA", 1, "filtered_quality", true, true, true, false, 9));
reads.Add(CreateRead("chr1", "AAA", 1, "Filtered_CigarData", true, true, true, false, 10, false));
reads.Add(CreateRead("chr1", "AAA", 1, "yay", true, true, true, false, 10));
var extractor = new MockAlignmentExtractor(reads);
var config = new AlignmentSourceConfig {
MinimumMapQuality = 10,
OnlyUseProperPairs = true, SkipDuplicates = true
};
var source = new AlignmentSource(extractor, null, config);
Read read;
var fetchedCount = 0;
while ((read = source.GetNextRead()) != null)
{
fetchedCount++;
Assert.Equal(read.Name, "yay");
}
Assert.Equal(1, fetchedCount);
// -------------------------
// turn off proper pairs
config.OnlyUseProperPairs = false;
fetchedCount = 0;
extractor.Reset();
source = new AlignmentSource(extractor, null, config);
while ((read = source.GetNextRead()) != null)
{
fetchedCount++;
Assert.Equal(read.Name,
fetchedCount == 1 ? "filtered_properpair" : "yay");
}
Assert.Equal(fetchedCount, 2);
// -------------------------
// change quality cut off
config.OnlyUseProperPairs = true;
config.MinimumMapQuality = 9;
fetchedCount = 0;
extractor.Reset();
source = new AlignmentSource(extractor, null, config);
while ((read = source.GetNextRead()) != null)
{
fetchedCount++;
Assert.Equal(read.Name,
fetchedCount == 1 ? "filtered_quality" : "yay");
}
Assert.Equal(fetchedCount, 2);
}
private ISomaticVariantCaller CreateMockVariantCaller(VcfFileWriter vcfWriter, ApplicationOptions options, ChrReference chrRef, MockAlignmentExtractor mae, IStrandBiasFileWriter biasFileWriter = null, string intervalFilePath = null)
{
var config = new AlignmentSourceConfig
{
MinimumMapQuality = options.MinimumMapQuality,
OnlyUseProperPairs = options.OnlyUseProperPairs,
};
IAlignmentStitcher stitcher = null;
if (options.StitchReads)
{
if (options.UseXCStitcher)
{
stitcher = new XCStitcher(options.MinimumBaseCallQuality);
}
else
{
stitcher = new BasicStitcher(options.MinimumBaseCallQuality);
}
}
var mateFinder = options.StitchReads ? new AlignmentMateFinder(MAX_FRAGMENT_SIZE) : null;
var RegionPadder = new RegionPadder(chrRef, null);
var alignmentSource = new AlignmentSource(mae, mateFinder, stitcher, config);
var variantFinder = new CandidateVariantFinder(options.MinimumBaseCallQuality, options.MaxSizeMNV, options.MaxGapBetweenMNV, options.CallMNVs);
var alleleCaller = new AlleleCaller(new VariantCallerConfig
{
IncludeReferenceCalls = options.OutputgVCFFiles,
MinVariantQscore = options.MinimumVariantQScore,
MaxVariantQscore = options.MaximumVariantQScore,
VariantQscoreFilterThreshold = options.FilteredVariantQScore > options.MinimumVariantQScore ? options.FilteredVariantQScore : (int?)null,
MinCoverage = options.MinimumCoverage,
MinFrequency = options.MinimumFrequency,
EstimatedBaseCallQuality = options.AppliedNoiseLevel == -1 ? options.MinimumBaseCallQuality : options.AppliedNoiseLevel,
StrandBiasModel = options.StrandBiasModel,
StrandBiasFilterThreshold = options.StrandBiasAcceptanceCriteria,
FilterSingleStrandVariants = options.FilterOutVariantsPresentOnlyOneStrand,
GenotypeModel = options.GTModel
});
var stateManager = new RegionStateManager();
return(new SomaticVariantCaller(
alignmentSource,
variantFinder,
alleleCaller,
vcfWriter,
stateManager,
chrRef,
RegionPadder,
biasFileWriter));
}