/// <summary>
/// Parse a single sequence using a MBFTextReader.
/// </summary>
/// <param name="bioText">sequence alignment text.</param>
/// <param name="isReadOnly">
/// Flag to indicate whether the sequences in the resulting sequence alignment should be in readonly mode or not.
/// If this flag is set to true then the resulting sequences's isReadOnly property
/// will be set to true, otherwise it will be set to false.
/// </param>
/// <param name="alphabet">Alphabet of the sequences.</param>
/// <param name="encoding">Required encoding.</param>
/// <param name="referenceSequences">Reference sequences.</param>
private static SAMAlignedSequence ParseSequence(string bioText, bool isReadOnly, IAlphabet alphabet, IEncoding encoding, IList <ISequence> referenceSequences)
{
const int optionalTokenStartingIndex = 11;
string[] tokens = bioText.Split(tabDelim, StringSplitOptions.RemoveEmptyEntries);
SAMAlignedSequence alignedSeq = new SAMAlignedSequence();
alignedSeq.QName = tokens[0];
alignedSeq.Flag = SAMAlignedSequenceHeader.GetFlag(tokens[1]);
alignedSeq.RName = tokens[2];
alignedSeq.Pos = int.Parse(tokens[3], CultureInfo.InvariantCulture);
alignedSeq.MapQ = int.Parse(tokens[4], CultureInfo.InvariantCulture);
alignedSeq.CIGAR = tokens[5];
alignedSeq.MRNM = tokens[6].Equals("=") ? alignedSeq.RName : tokens[6];
alignedSeq.MPos = int.Parse(tokens[7], CultureInfo.InvariantCulture);
alignedSeq.ISize = int.Parse(tokens[8], CultureInfo.InvariantCulture);
ISequence refSeq = null;
if (referenceSequences != null && referenceSequences.Count > 0)
{
refSeq = referenceSequences.FirstOrDefault(R => string.Compare(R.ID, alignedSeq.RName, StringComparison.OrdinalIgnoreCase) == 0);
}
ParseQualityNSequence(alignedSeq, alphabet, encoding, tokens[9], tokens[10], refSeq, isReadOnly);
SAMOptionalField optField = null;
string message;
for (int i = optionalTokenStartingIndex; i < tokens.Length; i++)
{
optField = new SAMOptionalField();
if (!Helper.IsValidRegexValue(OptionalFieldRegex, tokens[i]))
{
message = string.Format(CultureInfo.CurrentCulture, Resource.InvalidOptionalField, tokens[i]);
throw new FormatException(message);
}
string[] opttokens = tokens[i].Split(colonDelim, StringSplitOptions.RemoveEmptyEntries);
optField.Tag = opttokens[0];
optField.VType = opttokens[1];
optField.Value = opttokens[2];
alignedSeq.OptionalFields.Add(optField);
}
return(alignedSeq);
}
/// <summary>
/// Gets the paired reads type.
/// </summary>
/// <param name="read1">First aligned sequence.</param>
/// <param name="read2">Second aligned sequence.</param>
/// <param name="libraryName">library name.</param>
public static PairedReadType GetPairedReadType(SAMAlignedSequence read1, SAMAlignedSequence read2, string libraryName)
{
if (string.IsNullOrEmpty(libraryName))
{
throw new ArgumentNullException("libraryName");
}
CloneLibraryInformation libraryInfo = CloneLibrary.Instance.GetLibraryInformation(libraryName);
if (libraryInfo == null)
{
throw new ArgumentOutOfRangeException("libraryName");
}
return(GetPairedReadType(read1, read2, libraryInfo));
}
/// <summary>
/// Gets the insert length of reads.
/// </summary>
/// <param name="read1">First read.</param>
/// <param name="read2">Second read.</param>
/// <param name="validate">Validates the reads before calculating the insert length.</param>
public static int GetInsertLength(SAMAlignedSequence read1, SAMAlignedSequence read2, bool validate)
{
// reference chromosome
//5' --> 3'
//----------------------------------------------------- F strand
//
//3' <-- 5'
//----------------------------------------------------- R strand
// read1 read2
// 5' 3' 3' 5'
// --> <--
// |-------------- --------------|
// |<----------insert length------------------>|
if (read1 == null)
{
throw new ArgumentNullException("read1");
}
if (read2 == null)
{
return(0);
}
if (validate)
{
PairedReadType type = GetPairedReadType(read1, read2, 0, 0);
if (type != PairedReadType.Normal && type != PairedReadType.LengthAnomaly)
{
return(0);
}
}
if (read1.ISize == -read2.ISize)
{
return(read1.ISize >= 0 ? read1.ISize : -read1.ISize);
}
else
{
return(0);
}
}
/// <summary>
/// Parases sequence data and quality values and updates SAMAlignedSequence instance.
/// </summary>
/// <param name="alignedSeq">SAM aligned Sequence.</param>
/// <param name="alphabet">Alphabet of the sequence to be created.</param>
/// <param name="Encoding">Encoding to use while creating sequence.</param>
/// <param name="sequencedata">Sequence data.</param>
/// <param name="qualitydata">Quality values.</param>
/// <param name="refSeq">Reference sequence if known.</param>
/// <param name="isReadOnly">Flag to indicate whether the new sequence is required to in readonly or not.</param>
public static void ParseQualityNSequence(SAMAlignedSequence alignedSeq, IAlphabet alphabet, IEncoding Encoding, string sequencedata, string qualitydata, ISequence refSeq, bool isReadOnly)
{
if (alignedSeq == null)
{
throw new ArgumentNullException("alignedSeq");
}
if (string.IsNullOrWhiteSpace(sequencedata))
{
throw new ArgumentNullException("sequencedata");
}
if (string.IsNullOrWhiteSpace(qualitydata))
{
throw new ArgumentNullException("qualitydata");
}
bool isQualitativeSequence = true;
string message = string.Empty;
byte[] qualScores = null;
FastQFormatType fastQType = QualityFormatType;
if (sequencedata.Equals("*"))
{
return;
}
if (qualitydata.Equals("*"))
{
isQualitativeSequence = false;
}
if (isQualitativeSequence)
{
// Get the quality scores from the fourth line.
qualScores = ASCIIEncoding.ASCII.GetBytes(qualitydata);
// Check for sequence length and quality score length.
if (sequencedata.Length != qualitydata.Length)
{
string message1 = string.Format(CultureInfo.CurrentCulture, Resource.FastQ_InvalidQualityScoresLength, alignedSeq.QName);
message = string.Format(CultureInfo.CurrentCulture, Resource.IOFormatErrorMessage, Resource.SAM_NAME, message1);
Trace.Report(message);
throw new FileFormatException(message);
}
}
// get "." symbol indexes.
int index = sequencedata.IndexOf('.', 0);
while (index > -1)
{
alignedSeq.DotSymbolIndexes.Add(index++);
index = sequencedata.IndexOf('.', index);
}
// replace "." with N
if (alignedSeq.DotSymbolIndexes.Count > 0)
{
sequencedata = sequencedata.Replace('.', 'N');
}
// get "=" symbol indexes.
index = sequencedata.IndexOf('=', 0);
while (index > -1)
{
alignedSeq.EqualSymbolIndexes.Add(index++);
index = sequencedata.IndexOf('=', index);
}
// replace "=" with corresponding symbol from refSeq.
if (alignedSeq.EqualSymbolIndexes.Count > 0)
{
if (refSeq == null)
{
throw new ArgumentException(Resource.RefSequenceNofFound);
}
for (int i = 0; i < alignedSeq.EqualSymbolIndexes.Count; i++)
{
index = alignedSeq.EqualSymbolIndexes[i];
sequencedata = sequencedata.Remove(index, 1);
sequencedata = sequencedata.Insert(index, refSeq[index].Symbol.ToString());
}
}
ISequence sequence = null;
if (isQualitativeSequence)
{
QualitativeSequence qualSeq = null;
if (Encoding == null)
{
qualSeq = new QualitativeSequence(alphabet, fastQType, sequencedata, qualScores);
}
else
{
qualSeq = new QualitativeSequence(alphabet, fastQType, Encoding, sequencedata, qualScores);
}
qualSeq.ID = alignedSeq.QName;
qualSeq.IsReadOnly = isReadOnly;
sequence = qualSeq;
}
else
{
Sequence seq = null;
if (Encoding == null)
{
seq = new Sequence(alphabet, sequencedata);
}
else
{
seq = new Sequence(alphabet, Encoding, sequencedata);
}
seq.ID = alignedSeq.QName;
seq.IsReadOnly = isReadOnly;
sequence = seq;
}
alignedSeq.QuerySequence = sequence;
}
/// <summary>
/// Gets the paired reads when DV is enabled.
/// </summary>
/// <param name="meanLengthOfInsert">Mean of the insert length.</param>
/// <param name="standardDeviationOfInsert">Standard deviation of insert length.</param>
/// <param name="calculate">If this flag is set then mean and standard deviation will
/// be calculated from the paired reads instead of specified.</param>
/// <returns>List of paired read.</returns>
private IList <PairedRead> GetDVAwarePairedReads(float meanLengthOfInsert, float standardDeviationOfInsert, bool calculate = false)
{
// Dictionary helps to get the information at one pass of alinged sequence list.
Dictionary <string, DVEnabledPairedRead> pairedReads = new Dictionary <string, DVEnabledPairedRead>();
double sum = 0;
int count = 0;
for (int i = 0; i < QuerySequences.Count; i++)
{
DVEnabledPairedRead pairedRead;
SAMAlignedSequence read = QuerySequences[i];
if ((read.Flag & SAMFlags.PairedRead) == SAMFlags.PairedRead)
{
if (pairedReads.TryGetValue(read.QName, out pairedRead))
{
if (pairedRead.Index2 == -1 || pairedRead.Index1 == -1)
{
if (pairedRead.Index2 == -1)
{
pairedRead.Index2 = i;
}
else
{
pairedRead.Index1 = i;
}
// For best performace,
// 1. BAM/SAM file should be sorted by reads name.
// 2. If sorted on mapping position then give unmapped read a coordinate (generally the coordinate of the mapped mate)
// for sorting/indexing purposes only.
pairedRead.PairedType = PairedRead.GetPairedReadType(pairedRead.Read1, pairedRead.Read2, meanLengthOfInsert, standardDeviationOfInsert);
if (pairedRead.PairedType == PairedReadType.Normal || pairedRead.PairedType == PairedReadType.LengthAnomaly)
{
pairedRead.InsertLength = PairedRead.GetInsertLength(pairedRead.Read1, pairedRead.Read2);
if (calculate)
{
sum += pairedRead.InsertLength;
count++;
}
}
}
else
{
pairedRead.InsertLength = 0;
if (calculate)
{
sum -= pairedRead.InsertLength;
count--;
}
pairedRead.ReadIndexes.Add(i);
pairedRead.PairedType = PairedReadType.MultipleHits;
}
}
else
{
pairedRead = new DVEnabledPairedRead(QuerySequences);
if (!string.IsNullOrEmpty(read.RName) && !read.RName.Equals("*"))
{
pairedRead.Index1 = i;
}
else
{
pairedRead.Index2 = i;
}
pairedRead.PairedType = PairedReadType.Orphan;
pairedRead.InsertLength = 0;
pairedReads.Add(read.QName, pairedRead);
}
}
}
List <PairedRead> allreads = pairedReads.Values.ToList <PairedRead>();
pairedReads = null;
if (calculate && count > 0)
{
UpdateType(allreads, sum, count);
}
return(allreads);
}
/// <summary>
/// Gets the paired reads when SAMAligned sequences are in memory.
/// </summary>
/// <param name="meanLengthOfInsert">Mean of the insert length.</param>
/// <param name="standardDeviationOfInsert">Standard deviation of insert length.</param>
/// <param name="calculate">If this flag is set then mean and standard deviation will
/// be calculated from the paired reads instead of specified.</param>
/// <returns>List of paired read.</returns>
private IList <PairedRead> GetInMemoryPairedReads(float meanLengthOfInsert, float standardDeviationOfInsert, bool calculate = false)
{
// Dictionary helps to get the information at one pass of alinged sequence list.
Dictionary <string, PairedRead> pairedReads = new Dictionary <string, PairedRead>();
double sum = 0;
int count = 0;
for (int i = 0; i < QuerySequences.Count; i++)
{
PairedRead pairedRead;
SAMAlignedSequence read = QuerySequences[i];
if ((read.Flag & SAMFlags.PairedRead) == SAMFlags.PairedRead)
{
if (pairedReads.TryGetValue(read.QName, out pairedRead))
{
if (pairedRead.Read2 == null || pairedRead.Read1 == null)
{
if (pairedRead.Read2 == null)
{
pairedRead.Read2 = read;
}
else
{
pairedRead.Read1 = read;
}
pairedRead.PairedType = PairedRead.GetPairedReadType(pairedRead.Read1, pairedRead.Read2, meanLengthOfInsert, standardDeviationOfInsert);
if (pairedRead.PairedType == PairedReadType.Normal || pairedRead.PairedType == PairedReadType.LengthAnomaly)
{
pairedRead.InsertLength = PairedRead.GetInsertLength(pairedRead.Read1, pairedRead.Read2);
if (calculate)
{
sum += pairedRead.InsertLength;
count++;
}
}
}
else
{
pairedRead.InsertLength = 0;
if (calculate)
{
sum -= pairedRead.InsertLength;
count--;
}
pairedRead.Reads.Add(read);
pairedRead.PairedType = PairedReadType.MultipleHits;
}
}
else
{
pairedRead = new PairedRead();
if (!string.IsNullOrEmpty(read.RName) && !read.RName.Equals("*"))
{
pairedRead.Read1 = read;
}
else
{
pairedRead.Read2 = read;
}
pairedRead.PairedType = PairedReadType.Orphan;
pairedRead.InsertLength = 0;
pairedReads.Add(read.QName, pairedRead);
}
}
}
List <PairedRead> allreads = pairedReads.Values.ToList();
pairedReads = null;
if (calculate && count > 0)
{
UpdateType(allreads, sum, count);
}
return(allreads);
}
/// <summary>
/// Gets the insert length of reads.
/// </summary>
/// <param name="read1">First read.</param>
/// <param name="read2">Second read.</param>
public static int GetInsertLength(SAMAlignedSequence read1, SAMAlignedSequence read2)
{
return(GetInsertLength(read1, read2, false));
}
/// <summary>
/// Gets the paired reads type.
/// </summary>
/// <param name="read1">First aligned sequence.</param>
/// <param name="read2">Second aligned sequence.</param>
/// <param name="meanLengthOfInsert">Mean of the insertion length.</param>
/// <param name="standardDeviationOfInsert">Standard deviation of insertion length.</param>
public static PairedReadType GetPairedReadType(SAMAlignedSequence read1, SAMAlignedSequence read2, float meanLengthOfInsert, float standardDeviationOfInsert)
{
PairedReadType type = PairedReadType.Normal;
if (read1 == null)
{
throw new ArgumentNullException("read1");
}
if (read2 == null)
{
return(PairedReadType.Orphan);
}
if (string.IsNullOrEmpty(read2.RName) ||
read2.RName.Equals("*") ||
((read2.Flag & SAMFlags.UnmappedQuery) == SAMFlags.UnmappedQuery))
{
type = PairedReadType.Orphan;
}
else if (!read2.RName.Equals(read1.RName))
{
type = PairedReadType.Chimera;
}
else
{
bool isBothforwardReads = IsForwardRead(read1) && IsForwardRead(read2);
bool isBothReverseReads = IsReverseRead(read1) && IsReverseRead(read2);
if (isBothforwardReads || isBothReverseReads)
{
type = PairedReadType.StructuralAnomaly;
}
else
{
int forwardReadStartPos = 0;
int reverseReadStartPos = 0;
if (IsForwardRead(read1))
{
forwardReadStartPos = read1.Pos;
reverseReadStartPos = read2.Pos;
}
else
{
forwardReadStartPos = read2.Pos;
reverseReadStartPos = read1.Pos;
}
if (forwardReadStartPos > reverseReadStartPos)
{
type = PairedReadType.StructuralAnomaly;
}
else
{
int insertLength = GetInsertLength(read1, read2);
// µ + 3σ
float upperLimit = meanLengthOfInsert + (3 * standardDeviationOfInsert);
// µ - 3σ
float lowerLimit = meanLengthOfInsert - (3 * standardDeviationOfInsert);
if (insertLength > upperLimit || insertLength < lowerLimit)
{
type = PairedReadType.LengthAnomaly;
}
}
}
}
return(type);
}