public static bool HasInternalSoftclip(this CigarAlignment cigar)
{
var subCigar = cigar.GetSubCigar(cigar.GetPrefixClip() > 0 ? 1 : 0, cigar.Count - (cigar.GetSuffixClip() > 0 ? 1 : 0));
foreach (CigarOp op in subCigar)
{
if (op.Type == 'S')
{
return(true);
}
}
return(false);
}
public static AlignmentSummary GetAlignmentSummary(int startIndexInReference, CigarAlignment cigarData, string refSequence, string readSequence, bool trackActualMismatches = true, bool checkSoftclipsForMismatches = true, int probeSoftclipPrefix = 0, int probeSoftclipSuffix = 0)
{
var summary = new AlignmentSummary();
summary.Cigar = cigarData;
if (checkSoftclipsForMismatches)
{
startIndexInReference = startIndexInReference - (int)cigarData.GetPrefixClip();
}
var startIndexInRead = 0;
var anchorLength = 0;
var endAnchorLength = 0;
var hasHitNonMatch = false;
for (var cigarOpIndex = 0; cigarOpIndex < cigarData.Count; cigarOpIndex++)
{
var operation = cigarData[cigarOpIndex];
switch (operation.Type)
{
case 'S': // soft-clip
for (var i = 0; i < operation.Length; i++)
{
summary.NumSoftclips++;
if (readSequence[startIndexInRead + i] != 'N')
{
summary.NumNonNSoftclips++;
if (checkSoftclipsForMismatches)
{
if (startIndexInReference + i < 0 ||
startIndexInReference + i >= refSequence.Length)
{
summary.NumMismatchesIncludeSoftclip++;
}
else if (readSequence[startIndexInRead + i] !=
refSequence[startIndexInReference + i])
{
summary.NumMismatchesIncludeSoftclip++;
if (trackActualMismatches)
{
if (summary.MismatchesIncludeSoftclip == null)
{
summary.MismatchesIncludeSoftclip = new List <string> {
};
}
var mismatch = string.Format("{0}_{1}_{2}",
startIndexInReference + i,
refSequence[startIndexInReference + i],
readSequence[startIndexInRead + i]);
summary.MismatchesIncludeSoftclip.Add(mismatch);
}
}
}
}
}
break;
case 'M': // match or mismatch
for (var i = 0; i < operation.Length; i++)
{
if (startIndexInReference + i > refSequence.Length - 1)
{
return(null);
throw new InvalidDataException(
"Read goes off the end of the genome: " + startIndexInReference + ":" +
cigarData.ToString() + " vs " + startIndexInReference + " + " + refSequence.Length);
}
var baseAtIndex = readSequence[startIndexInRead + i];
if (baseAtIndex != 'N' && baseAtIndex !=
refSequence[startIndexInReference + i])
{
summary.NumMismatches++;
summary.NumMismatchesIncludeSoftclip++;
if (trackActualMismatches)
{
if (summary.MismatchesIncludeSoftclip == null)
{
summary.MismatchesIncludeSoftclip = new List <string> {
};
}
var mismatch = string.Format("{0}_{1}_{2}", startIndexInReference + i,
refSequence[startIndexInReference + i], readSequence[startIndexInRead + i]);
summary.MismatchesIncludeSoftclip.Add(mismatch);
}
hasHitNonMatch = true;
endAnchorLength = 0;
}
else
{
if (baseAtIndex != 'N')
{
summary.NumMatches++;
}
if (!hasHitNonMatch)
{
anchorLength++;
}
endAnchorLength++;
}
}
break;
case 'I': // insertion
hasHitNonMatch = true;
endAnchorLength = 0;
summary.NumIndels++;
summary.NumIndelBases += (int)operation.Length;
summary.NumInsertedBases += (int)operation.Length;
break;
case 'D': // deletion
hasHitNonMatch = true;
endAnchorLength = 0;
summary.NumIndels++;
summary.NumIndelBases += (int)operation.Length;
summary.NumDeletedBases += (int)operation.Length;
break;
}
if (operation.IsReadSpan())
{
startIndexInRead += (int)operation.Length;
}
if (operation.IsReferenceSpan())
{
startIndexInReference += (int)operation.Length;
}
if (checkSoftclipsForMismatches && operation.Type == 'S')
{
startIndexInReference += (int)operation.Length;
}
}
summary.AnchorLength = Math.Min(anchorLength, endAnchorLength);
return(summary);
}
public StitchingInfo GetStitchedCigar(CigarAlignment cigar1, int pos1, CigarAlignment cigar2, int pos2, bool reverseFirst, bool pairIsOutie)
{
var positions = GetStitchedSites(cigar1, cigar2, pos2, pos1);
var success = true;
var stitchingInfo = ReconcileSites(positions, reverseFirst, out success, pairIsOutie ? (int)cigar2.GetPrefixClip() : (int)cigar1.GetPrefixClip(), pairIsOutie ? (int)(cigar1.GetReadSpan() - (int)cigar1.GetSuffixClip()) : (int)(cigar2.GetReadSpan() - (int)cigar2.GetSuffixClip()), pairIsOutie);
return(success ? stitchingInfo : null);
}
/// <summary>
/// Log result info to a result file. Doesn't directly impact test, just useful for looking at the results all together.
/// (Also useful for a deliverable output summary).
/// </summary>
/// <param name="resultFile"></param>
/// <param name="scenario"></param>
/// <param name="didStitch"></param>
/// <param name="resultSet"></param>
private void LogResult(string resultFile, StitchingScenario scenario, bool didStitch, AlignmentSet resultSet = null, string message = null)
{
var diagramLength = 12;
// This is useful for looking at the results across the full test set.
const string delimiter = ",";
var visualResultsFile = resultFile + ".visuals.csv";
if (!File.Exists(visualResultsFile))
{
using (var sw = File.CreateText(visualResultsFile))
{
var leftOfDiagram = new List <string>()
{
"ID", "Pos", "Cigar", "Dirs", "Diagram Var"
};
var varDiagram = Enumerable.Repeat("", diagramLength);
var leftOfRef = new List <string>()
{
"Pos", "Cigar", "Dirs", "Diagram Ref"
};
var refDiagram = Enumerable.Repeat("", diagramLength);
var leftOfStitched = new List <string>()
{
"Pos", "Cigar", "Dirs", "Diagram Stitched"
};
var stitchedDiagram = Enumerable.Repeat("", diagramLength);
sw.WriteLine(string.Join(delimiter,
leftOfDiagram.
Concat(varDiagram).
Concat(leftOfRef).
Concat(refDiagram).
Concat(leftOfStitched).
Concat(stitchedDiagram)));
}
}
using (var sw = File.AppendText(visualResultsFile))
{
Read stitchedRead = null;
if (resultSet != null && resultSet.ReadsForProcessing.Any())
{
stitchedRead = resultSet.ReadsForProcessing.First();
}
// First row
var leftOfDiagram = new List <string>()
{
scenario.Category + "-" + scenario.Id,
scenario.InputRead1.Position.ToString(), scenario.InputRead1.Cigar, scenario.InputRead1.Directions
};
var r1Cigar = new CigarAlignment(scenario.InputRead1.Cigar);
var r2Cigar = new CigarAlignment(scenario.InputRead2.Cigar);
var r1BasesStart = scenario.InputRead1.Position - 1 - (int)r1Cigar.GetPrefixClip();
var r2BasesStart = scenario.InputRead2.Position - 1 - (int)r2Cigar.GetPrefixClip();
if (r1BasesStart < 0 || r2BasesStart < 0)
{
throw new ArgumentException("Test scenario has invalid position/cigar combination: " + scenario.InputRead1.Position + ":" + scenario.InputRead1.Cigar + " or " + scenario.InputRead2.Position + ":" + scenario.InputRead2.Cigar);
}
if (r1BasesStart < 0)
{
r1BasesStart = 0;
}
if (r2BasesStart < 0)
{
r2BasesStart = 0;
}
var r2CigarLength = 0;
foreach (CigarOp op in r2Cigar)
{
r2CigarLength += (int)op.Length;
}
var r2BasesEnd = r2BasesStart + r2CigarLength;
var preOverlapCigar = new CigarAlignment(scenario.InputRead1.Cigar).GetClippedCigar(0, (int)(r2BasesStart - r1BasesStart) + 1, includeWholeEndIns: true);
var insertionsPreOverlap = preOverlapCigar.CountOperations('I');
var expectedReadLength = r2BasesEnd - r1BasesStart + insertionsPreOverlap;
r2BasesStart = r2BasesStart + insertionsPreOverlap;
var varDiagram = Enumerable.Repeat("", r1BasesStart).Concat(ExpandCigar(r1Cigar,
new CigarDirection(scenario.InputRead1.Directions))).ToList();
varDiagram = varDiagram.Concat(Enumerable.Repeat("", diagramLength - varDiagram.Count()).ToList()).ToList();
var leftOfRef = new List <string>()
{
scenario.InputRead2.Position.ToString(), scenario.InputRead2.Cigar, scenario.InputRead2.Directions
};
var refDiagram = Enumerable.Repeat("", diagramLength);
var leftOfStitched = Enumerable.Repeat("", 3).ToList();
var stitchedDiagram = Enumerable.Repeat("NA", diagramLength);
if (stitchedRead != null && stitchedRead.CigarDirections != null)
{
var stitchedBasesStart = stitchedRead.Position - 1 - (int)stitchedRead.CigarData.GetPrefixClip();
leftOfStitched = new List <string>()
{
stitchedRead.Position.ToString(), stitchedRead.CigarData.ToString(), GetDirectionsString(stitchedRead)
};
stitchedDiagram = Enumerable.Repeat("", stitchedBasesStart).Concat(ExpandCigar(stitchedRead.CigarData,
stitchedRead.CigarDirections));
}
sw.WriteLine(string.Join(delimiter,
leftOfDiagram.
Concat(varDiagram).
Concat(leftOfRef).
Concat(refDiagram).
Concat(leftOfStitched).
Concat(stitchedDiagram)));
// Second row
var varDiagramR2 =
Enumerable.Repeat("", r2BasesStart)
.Concat(ExpandCigar(r2Cigar,
new CigarDirection(scenario.InputRead2.Directions)));
varDiagramR2 = varDiagramR2.Concat(Enumerable.Repeat("", diagramLength - varDiagramR2.Count())).ToList();
var leftOfDiagramPad = new List <string>()
{
"",
scenario.InputRead2.Position.ToString(),
scenario.InputRead2.Cigar,
scenario.InputRead2.Directions
};
var leftOfRefPad = Enumerable.Repeat("", leftOfRef.Count);
var leftOfStitchedPad = Enumerable.Repeat("", leftOfStitched.Count);
var refDiagramR2 = Enumerable.Repeat("", diagramLength);
var totalBasesCovered =
Enumerable.Repeat("", r1BasesStart).Concat(Enumerable.Repeat("+", expectedReadLength)).ToList();
sw.WriteLine(string.Join(delimiter,
leftOfDiagramPad.
Concat(varDiagramR2).
Concat(leftOfRefPad).
Concat(refDiagramR2).
Concat(leftOfStitchedPad).
Concat(totalBasesCovered)
));
sw.WriteLine();
}
if (!File.Exists(resultFile))
{
// Create a file to write to, and write the header.
using (var sw = File.CreateText(resultFile))
{
sw.WriteLine(string.Join(delimiter, new[] { "ID",
"R1_Pos", "R1_Cigar", "R1_Dirs",
"R2_Pos", "R2_Cigar", "R2_Dirs",
"ShouldStitch", "DidStitch",
"Exp_SR_Pos", "Exp_SR_Cigar", "Exp_SR_Dirs",
"Actual_SR_Pos", "Actual_SR_Cigar", "Actual_SR_Dirs",
"Notes", "Pass", "Message" }));
}
}
using (var sw = File.AppendText(resultFile))
{
// Add everything we know from the input scenario, and whether it did stitch.
var fields = new List <string>()
{
scenario.Category + "-" + scenario.Id,
scenario.InputRead1.Position.ToString(), scenario.InputRead1.Cigar, scenario.InputRead1.Directions,
scenario.InputRead2.Position.ToString(), scenario.InputRead2.Cigar, scenario.InputRead2.Directions,
scenario.ShouldStitch.ToString(), didStitch.ToString(),
scenario.OutputRead1.Position.ToString(), scenario.OutputRead1.Cigar, scenario.OutputRead1.Directions,
};
var stitchResultsMatch = false;
var cigarResultsMatch = false;
var directionResultsMatch = false;
stitchResultsMatch = scenario.ShouldStitch == didStitch;
// Add the info from the output reads
if (resultSet != null && resultSet.ReadsForProcessing.Any() && resultSet.ReadsForProcessing.First().CigarDirections != null)
{
var stitchedRead = resultSet.ReadsForProcessing.First();
var directions = GetDirectionsString(stitchedRead);
fields.AddRange(new List <string>()
{
stitchedRead.Position.ToString(),
stitchedRead.CigarData.ToString(),
directions
});
cigarResultsMatch = !scenario.ShouldStitch || OutputCigarsMatch(scenario, resultSet);
directionResultsMatch = !scenario.ShouldStitch || OutputDirectionsMatch(scenario, resultSet);
}
else
{
fields.AddRange(new List <string>()
{
"", "", ""
});
}
// Determine if this scenario "Passed" (i.e. matched expectations). (TODO if the resultSet is null, it failed -- is that valid?)
var testResult = (!scenario.ShouldStitch && stitchResultsMatch) || (stitchResultsMatch && cigarResultsMatch && directionResultsMatch);
fields.Add(Sanitize(scenario.Notes, Convert.ToChar(delimiter)));
fields.Add(testResult.ToString());
fields.Add(message);
// Write scenario results to file
sw.WriteLine(string.Join(delimiter, fields));
}
}
public static CigarAlignment GetCigarWithoutProbeClips(this CigarAlignment cigar, bool isRead1)
{
return(isRead1 ?
cigar.GetSubCigar(cigar.GetPrefixClip() > 0 ? 1 : 0, cigar.Count) :
cigar.GetSubCigar(0, cigar.Count - (cigar.GetSuffixClip() > 0 ? 1 : 0)));
}
public static uint GetReadSpanBetweenClippedEnds(this CigarAlignment cigar)
{
return(cigar.GetReadSpan() - cigar.GetPrefixClip() - cigar.GetSuffixClip());
}
public static AlignmentSummary GetAlignmentSummary(int startIndexInReference, CigarAlignment cigarData, string refSequence, string readSequence, bool trackActualMismatches = true, bool checkSoftclipsForMismatches = true, int probeSoftclipPrefix = 0, int probeSoftclipSuffix = 0)
{
var summary = new AlignmentSummary();
summary.Cigar = cigarData;
if (checkSoftclipsForMismatches)
{
startIndexInReference = startIndexInReference - (int)cigarData.GetPrefixClip();
}
var startIndexInRead = 0;
var anchorLength = 0;
var endAnchorLength = 0;
var hasHitNonMatch = false;
var hasHitNonNSoftclip = false;
for (var cigarOpIndex = 0; cigarOpIndex < cigarData.Count; cigarOpIndex++)
{
var operation = cigarData[cigarOpIndex];
var opLength = (int)(operation.Length);
switch (operation.Type)
{
case 'S': // soft-clip
for (var i = 0; i < opLength; i++)
{
summary.NumSoftclips++;
// No special treatement for Ns that are inside the softclip. Because the whole N-softclip distinction was meant to deal with padding-type softclips, I think.
if (readSequence[startIndexInRead + i] != 'N' || hasHitNonNSoftclip)
{
hasHitNonNSoftclip = true;
summary.NumNonNSoftclips++;
if (checkSoftclipsForMismatches)
{
if (startIndexInReference + i < 0 ||
startIndexInReference + i >= refSequence.Length)
{
summary.NumMismatchesIncludeSoftclip++;
}
else if (readSequence[startIndexInRead + i] !=
refSequence[startIndexInReference + i] && readSequence[startIndexInRead + i] != 'N')
{
summary.NumMismatchesIncludeSoftclip++;
if (trackActualMismatches)
{
if (summary.MismatchesIncludeSoftclip == null)
{
summary.MismatchesIncludeSoftclip = new List <string> {
};
}
// TODO WHEN KILL HYGEA, remove this if we're not using anymore, to save time
var mismatch = string.Format("{0}_{1}_{2}",
startIndexInReference + i,
refSequence[startIndexInReference + i],
readSequence[startIndexInRead + i]);
summary.MismatchesIncludeSoftclip.Add(mismatch);
}
}
}
}
//else
//{
// if (!hasHitNonNSoftclip)
// {
// nSoftclipLength++;
// }
//}
}
break;
case 'M': // match or mismatch
for (var i = 0; i < opLength; i++)
{
if (startIndexInReference + i > refSequence.Length - 1)
{
return(null);
throw new InvalidDataException(
"Read goes off the end of the genome: " + startIndexInReference + ":" +
cigarData.ToString() + " vs " + startIndexInReference + " + " + refSequence.Length);
}
if (startIndexInReference + i < 0)
{
throw new InvalidDataException(
"Read would be before beginning of the chromosome: " + startIndexInReference + ":" +
cigarData.ToString() + " vs " + startIndexInReference + " + " + refSequence.Length);
}
var baseAtIndex = readSequence[startIndexInRead + i];
if (baseAtIndex != 'N' && baseAtIndex !=
refSequence[startIndexInReference + i])
{
summary.NumMismatches++;
summary.NumMismatchesIncludeSoftclip++;
if (trackActualMismatches)
{
if (summary.MismatchesIncludeSoftclip == null)
{
summary.MismatchesIncludeSoftclip = new List <string> {
};
}
// TODO WHEN KILL HYGEA, remove this if we're not using anymore, to save time
var mismatch = string.Format("{0}_{1}_{2}", startIndexInReference + i,
refSequence[startIndexInReference + i], readSequence[startIndexInRead + i]);
summary.MismatchesIncludeSoftclip.Add(mismatch);
}
hasHitNonMatch = true;
endAnchorLength = 0;
}
else
{
if (baseAtIndex != 'N')
{
summary.NumMatches++;
}
if (!hasHitNonMatch)
{
anchorLength++;
}
endAnchorLength++;
}
}
break;
case 'I': // insertion
hasHitNonMatch = true;
endAnchorLength = 0;
summary.NumIndels++;
summary.NumIndelBases += opLength;
summary.NumInsertedBases += opLength;
break;
case 'D': // deletion
hasHitNonMatch = true;
endAnchorLength = 0;
summary.NumIndels++;
summary.NumIndelBases += opLength;
summary.NumDeletedBases += opLength;
break;
}
if (operation.IsReadSpan())
{
startIndexInRead += opLength;
}
if (operation.IsReferenceSpan())
{
startIndexInReference += opLength;
}
if (checkSoftclipsForMismatches && operation.Type == 'S')
{
startIndexInReference += opLength;
}
}
summary.AnchorLength = Math.Min(anchorLength, endAnchorLength);
return(summary);
}
