/// <summary>
/// Gets a boolean value which indicates that whether the specified read is reverse oriented or not.
/// </summary>
/// <param name="read">Aligned Sequence.</param>
public static bool IsReverseRead(SAMAlignedSequence read)
{
if (read == null)
{
throw new ArgumentNullException("read");
}
return(!IsForwardRead(read));
}
/// <summary>
/// Gets the paired reads type.
/// </summary>
/// <param name="read1">First aligned sequence.</param>
/// <param name="read2">Second aligned sequence.</param>
/// <param name="libraryInfo">Library information.</param>
public static PairedReadType GetPairedReadType(SAMAlignedSequence read1, SAMAlignedSequence read2, CloneLibraryInformation libraryInfo)
{
if (libraryInfo == null)
{
throw new ArgumentNullException("libraryInfo");
}
return(GetPairedReadType(read1, read2, libraryInfo.MeanLengthOfInsert, libraryInfo.StandardDeviationOfInsert));
}
/// <summary>
/// Gets a boolean value which indicates that whether the specified read is forward oriented or not.
/// </summary>
/// <param name="read">Aligned Sequence.</param>
public static bool IsForwardRead(SAMAlignedSequence read)
{
if (read == null)
{
throw new ArgumentNullException("read");
}
return((read.Flag & SAMFlags.QueryOnReverseStrand) == 0);
}
public static string GetOptionValue(this SAMAlignedSequence sam, string tag, string vtype, string defaultValue)
{
var field = sam.OptionalFields.FirstOrDefault(m => m.Tag.Equals(tag) && m.VType.Equals(vtype));
if (null == field)
{
return(defaultValue);
}
return(field.Value);
}
public static string GetQualityScoresString(this SAMAlignedSequence sam)
{
if (sam.Flag.HasFlag(SAMFlags.QueryOnReverseStrand))
{
return(new string(sam.GetEncodedQualityScores().Reverse().Select(a => (char)a).ToArray()));
}
else
{
return(new string(sam.GetEncodedQualityScores().Select(a => (char)a).ToArray()));
}
}
public static string GetQuerySequenceString(this SAMAlignedSequence sam)
{
if (sam.Flag.HasFlag(SAMFlags.QueryOnReverseStrand))
{
return(sam.QuerySequence.GetReverseComplementedSequence().GetSequenceString());
}
else
{
return(sam.QuerySequence.GetSequenceString());
}
}
/// <summary>
/// Parses 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="sequencedata">Sequence data.</param>
/// <param name="qualitydata">Quality values.</param>
/// <param name="validate">Validation needed</param>
public static void ParseQualityNSequence(SAMAlignedSequence alignedSeq, IAlphabet alphabet, byte[] sequencedata, byte[] qualitydata, bool validate = true)
{
if (alignedSeq == null)
{
throw new ArgumentNullException("alignedSeq");
}
if (sequencedata == null || sequencedata.Length == 0)
{
throw new ArgumentNullException("sequencedata");
}
if (qualitydata == null || qualitydata.Length == 0)
{
throw new ArgumentNullException("qualitydata");
}
bool isQualitativeSequence = true;
string message = string.Empty;
FastQFormatType fastQType = QualityFormatType;
if (sequencedata.Length == 1 && sequencedata[0] == AsteriskAsByte)
{
return;
}
if (qualitydata.Length == 1 && qualitydata[0] == AsteriskAsByte)
{
isQualitativeSequence = false;
}
if (isQualitativeSequence)
{
// Check for sequence length and quality score length.
if (sequencedata.Length != qualitydata.Length)
{
string message1 = string.Format(CultureInfo.CurrentCulture, Properties.Resource.FastQ_InvalidQualityScoresLength, alignedSeq.QName);
message = string.Format(CultureInfo.CurrentCulture, Properties.Resource.IOFormatErrorMessage, Properties.Resource.SAM_NAME, message1);
Trace.Report(message);
throw new Exception(message);
}
}
alignedSeq.QuerySequence = isQualitativeSequence
? (ISequence) new QualitativeSequence(alphabet, fastQType, sequencedata, qualitydata, validate)
{
ID = alignedSeq.QName
}
: new Sequence(alphabet, sequencedata, validate)
{
ID = alignedSeq.QName
};
}
/// <summary>
/// Parses alignments in SAM format from a reader into a SequenceAlignmentMap object.
/// </summary>
/// <param name="reader">A reader for a biological sequence alignment text.</param>
/// <returns>A new SequenceAlignmentMap instance containing parsed data.</returns>
public IEnumerable <SAMAlignedSequence> ParseSequencesAsEnumerable(string fileName)
{
FileInfo fileInfo = new FileInfo(fileName);
using (StreamReader reader = new StreamReader(fileName))
{
if (reader == null)
{
throw new ArgumentNullException("reader");
}
// Parse the header lines and store them in a string.
// This is being done as parsing the header using the textreader is parsing an extra line.
List <string> headerStrings = new List <string>();
string line = ReadNextLine(reader);
while (line != null && line.StartsWith(@"@", StringComparison.OrdinalIgnoreCase))
{
headerStrings.Add(line);
line = ReadNextLine(reader);
}
// Parse the alignment header strings.
SAMAlignmentHeader header = ParseSamHeader(headerStrings);
SequenceAlignmentMap sequenceAlignmentMap = new SequenceAlignmentMap(header);
List <string> refSeqNames = null;
bool hasSQHeader = header.ReferenceSequences.Count > 0;
if (!hasSQHeader)
{
refSeqNames = new List <string>();
}
// Parse aligned sequences
// If the SQ header is not present in header then get the reference sequences information from reads.
while (line != null && !line.StartsWith(@"@", StringComparison.OrdinalIgnoreCase))
{
SAMAlignedSequence alignedSeq = ParseSequence(line, this.Alphabet);
if (!hasSQHeader)
{
if (!alignedSeq.RName.Equals("*", StringComparison.OrdinalIgnoreCase) &&
!refSeqNames.Contains(alignedSeq.RName, StringComparer.OrdinalIgnoreCase))
{
refSeqNames.Add(alignedSeq.RName);
}
}
yield return(alignedSeq);
//sequenceAlignmentMap.QuerySequences.Add(alignedSeq);
line = ReadNextLine(reader);
}
}
}
//public static string GetZezValue(this SAMAlignedSequence sam)
//{
// return GetOptionValue(sam, "ZE", "Z");
//}
public static void WriteFastq(this SAMAlignedSequence sam, StreamWriter sw, bool posAsPaired = false)
{
if (posAsPaired)
{
sw.WriteLine(string.Format("@{0} {1}", sam.QName, sam.Pos));
}
else
{
sw.WriteLine("@" + sam.QName);
}
sw.WriteLine(sam.GetQuerySequenceString());
sw.WriteLine("+");
sw.WriteLine(sam.GetQualityScoresString());
}
/// <summary>
/// Parse a single sequencer.
/// </summary>
/// <param name="bioText">sequence alignment text.</param>
/// <param name="alphabet">Alphabet of the sequences.</param>
/// <param name="referenceSequences">Reference sequences.</param>
private static SAMAlignedSequence ParseSequence(string bioText, IAlphabet alphabet, 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, tokens[9], tokens[10], refSeq);
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, Properties.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);
}
}
public static string GetOptionValue(this SAMAlignedSequence sam, string tag, string vtype, bool throwException = true, string parserName = null)
{
var field = sam.OptionalFields.FirstOrDefault(m => m.Tag.Equals(tag) && m.VType.Equals(vtype));
if (null == field)
{
if (throwException)
{
throw new Exception(string.Format("data error, cannot find {0}:{1}:XXX value in query {2}{3}.",
tag,
vtype,
sam.QName,
parserName == null ? "" : " by parser " + parserName));
}
else
{
return(null);
}
}
return(field.Value);
}
public IEnumerable<ISequence> Parse(Stream stream)
{
FastQParser fqp = new FastQParser ();
foreach (var seq in fqp.Parse (stream)) {
var name = seq.ID;
var sp = name.Split ('/');
var movie = sp [0];
var hole = sp [1];
SAMAlignedSequence sam = new SAMAlignedSequence ();
sam.QuerySequence = seq;
sam.OptionalFields.Add (new SAMOptionalField () { Tag = "sn", Value = "f,0,0,0,0" });
sam.OptionalFields.Add (new SAMOptionalField () { Tag = "rs", Value = "f,0,0,0,0,0,0" });
sam.OptionalFields.Add (new SAMOptionalField () { Tag = "zs", Value = "f,0,0,0,0,0,0" });
PacBioCCSRead read = new PacBioCCSRead (sam) {
AvgZscore = Single.NaN,
HoleNumber = Convert.ToInt32 (hole),
Movie = movie
};
yield return read;
}
}
/// <summary>
/// Parse a single sequencer.
/// </summary>
/// <param name="bioText">sequence alignment text.</param>
/// <param name="alphabet">Alphabet of the sequences.</param>
public static SAMAlignedSequence ParseSequence(string bioText, IAlphabet alphabet)
{
const int optionalTokenStartingIndex = 11;
string[] tokens = bioText.Split(TabDelim, StringSplitOptions.RemoveEmptyEntries);
SAMAlignedSequence alignedSeq = new SAMAlignedSequence
{
QName = tokens[0],
Flag = SAMAlignedSequenceHeader.GetFlag(tokens[1]),
RName = tokens[2],
Pos = int.Parse(tokens[3]),
MapQ = int.Parse(tokens[4]),
CIGAR = tokens[5]
};
alignedSeq.MRNM = tokens[6].Equals("=") ? alignedSeq.RName : tokens[6];
alignedSeq.MPos = int.Parse(tokens[7]);
alignedSeq.ISize = int.Parse(tokens[8]);
ParseQualityNSequence(alignedSeq, alphabet, tokens[9], tokens[10]);
for (int i = optionalTokenStartingIndex; i < tokens.Length; i++)
{
SAMOptionalField optField = new SAMOptionalField();
if (!Helper.IsValidRegexValue(OptionalFieldRegex, tokens[i]))
{
throw new FormatException(string.Format(Properties.Resource.InvalidOptionalField, tokens[i]));
}
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);
}
public string SAMToString(SAMAlignedSequence sam)
{
if (sam == null)
{
return null;
}
return string.Format("{0}\t{1}\t{2}\t{3}\t{4}\t{5}\t{6}\t{7}\t{8}\t{9}\t{10}\t{11}",
sam.QName,
(int)sam.Flag,
sam.RName,
sam.Pos,
sam.MapQ,
sam.CIGAR,
sam.MRNM,
sam.MPos,
sam.ISize,
sam.GetQuerySequenceString(),
sam.GetQualityScoresString(),
(from of in sam.OptionalFields
select string.Format("{0}:{1}:{2}", of.Tag, of.VType, of.Value)).Merge("\t"));
}
/// <summary>
/// Gets encoded sequence according to the BAM specification.
/// </summary>
/// <param name="alignedSeq"></param>
/// <returns></returns>
private static byte[] GetEncodedSequence(SAMAlignedSequence alignedSeq)
{
List<byte> byteList = new List<byte>();
ISequence seq = alignedSeq.QuerySequence;
if (seq != null)
{
if (!(seq.Alphabet is DnaAlphabet))
{
throw new ArgumentException(Properties.Resource.BAMFormatterSupportsDNAOnly);
}
byte[] symbolMap = seq.Alphabet.GetSymbolValueMap();
for (int i = 0; i < seq.Count; i++)
{
char symbol = (char)symbolMap[seq[i]];
byte encodedvalue = 0;
// 4-bit encoded read: =ACMGRSVTWYHKDBN -> 0-15; the earlier base is stored in the
// high-order 4 bits of the byte.
//Note:
// All the other symbols which are not supported by BAM specification (other than "=ACMGRSVTWYHKDBN") are converted to 'N'
// for example a '.' symbol which is supported by SAM specification will be converted to symbol 'N'
switch (symbol)
{
case '=':
encodedvalue = 0;
break;
case 'A':
encodedvalue = 1;
break;
case 'C':
encodedvalue = 2;
break;
case 'M':
encodedvalue = 3;
break;
case 'G':
encodedvalue = 4;
break;
case 'R':
encodedvalue = 5;
break;
case 'S':
encodedvalue = 6;
break;
case 'V':
encodedvalue = 7;
break;
case 'T':
encodedvalue = 8;
break;
case 'W':
encodedvalue = 9;
break;
case 'Y':
encodedvalue = 10;
break;
case 'H':
encodedvalue = 11;
break;
case 'K':
encodedvalue = 12;
break;
case 'D':
encodedvalue = 13;
break;
case 'B':
encodedvalue = 14;
break;
default:
encodedvalue = 15;
break;
}
if ((i + 1) % 2 > 0)
{
byteList.Add((byte)(encodedvalue << 4));
}
else
{
byteList[byteList.Count - 1] = (byte)(byteList[byteList.Count - 1] | encodedvalue);
}
}
}
return byteList.ToArray();
}
// Gets block size required for the specified SAMAlignedSequence object.
private int GetBlockSize(SAMAlignedSequence alignedSeq)
{
int readNameLen = alignedSeq.QName.Length + 1;
int cigarLen = GetCIGARLength(alignedSeq.CIGAR);
int readLen = (int)alignedSeq.QuerySequence.Count;
return 32 + readNameLen + (cigarLen * 4) + ((readLen + 1) / 2) + readLen + GetAuxiliaryDataLength(alignedSeq);
}
List<BaseAndQualityAndPosition> getBasesForSequence(SAMAlignedSequence seq)
{
List<BaseAndQualityAndPosition> toReturn = new List<BaseAndQualityAndPosition>(seq.RefEndPos - seq.Pos + 10);
// Decode the cigar string into operations.
// TODO: This code is duplicated in many places
string CIGAR = seq.CIGAR;
List<KeyValuePair<char, int>> charsAndPositions = new List<KeyValuePair<char, int>>();
for (int i = 0; i < CIGAR.Length; i++)
{
char ch = CIGAR[i];
if (Char.IsDigit(ch))
{
continue;
}
charsAndPositions.Add(new KeyValuePair<char, int>(ch, i));
}
// Get sequence bases and error probabilities
var qseq = seq.QuerySequence as QualitativeSequence;
var seq_log10ErrorProb = qseq.GetPhredQualityScores().Select(Utils.GetLog10ErrorProbability).ToArray();
var seq_bases = qseq.ToArray();
// Use the cigar operations to emit bases.
int curRef = seq.Pos;
int curQuery = 0;
for (int i = 0; i < charsAndPositions.Count; i++)
{
// Parse the current cigar operation
char ch = charsAndPositions[i].Key;
int cig_start = i==0 ? 0 : charsAndPositions[i - 1].Value + 1;
int cig_end = charsAndPositions[i].Value - cig_start;
int cig_len = int.Parse(CIGAR.Substring(cig_start, cig_end));
// Emit or advance based on cigar operation.
switch (ch)
{
case 'P': //padding (Silent deltions from padded reference)
case 'N': //skipped region from reference
throw new Exception("Pile up methods not built to handle reference clipping (Cigar P or N) yet.");
case 'M': //match or mismatch
case '=': //match
case 'X': //mismatch
for (int k = 0; k < cig_len; k++)
{
var bqp= new BaseAndQualityAndPosition(curRef,0, new BaseAndQuality(seq_bases[curQuery], seq_log10ErrorProb[curQuery]));
toReturn.Add(bqp);
curQuery++;
curRef++;
}
break;
case 'I'://insertion to the reference
for (int k = 0; k < cig_len; k++)
{
var bqp = new BaseAndQualityAndPosition(curRef,k, new BaseAndQuality(seq_bases[curQuery], seq_log10ErrorProb[curQuery]));
toReturn.Add(bqp);
curQuery++;
}
break;
case 'D'://Deletion from the reference
for (int k = 0; k < cig_len; k++)
{
var bqp = new BaseAndQualityAndPosition(curRef,k, new BaseAndQuality((byte)'-', Double.NaN));
toReturn.Add(bqp);
curRef++;
}
break;
case 'S': //soft clipped
curQuery += cig_len;
break;
case 'H'://had clipped
break;
default:
throw new FormatException("Unexpected SAM Cigar element found " + ch.ToString());
}
}
return toReturn;
}
/// <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>
/// 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>
/// Filters Sequence based on user inputs.
/// </summary>
/// <param name="alignedSequence">Aligned Sequence.</param>
/// <returns>Whether aligned sequence matches user defined options.</returns>
private bool Filter(SAMAlignedSequence alignedSequence)
{
bool filter = true;
if (filter && FlagRequired != 0)
{
filter = (((int)alignedSequence.Flag) & FlagRequired) == FlagRequired;
}
if (filter && FilteringFlag != 0)
{
filter = ((((int)alignedSequence.Flag) & FilteringFlag) == 0);
}
if (filter && QualityMinimumMapping != 0)
{
filter = alignedSequence.MapQ == QualityMinimumMapping;
}
if (filter && !string.IsNullOrEmpty(Library) && rgRecFields.Count > 0)
{
filter = rgRecFields.First(
a => a.Tags.First(
b => b.Tag.Equals("ID")).Value.Equals(alignedSequence.OptionalFields.First(
c => c.Tag.Equals("RG")).Value)).Tags.First(
d => d.Tag.Equals("LB")).Value.Equals(Library);
}
if (filter && !string.IsNullOrEmpty(ReadGroup))
{
filter = alignedSequence.OptionalFields.AsParallel().Where(
O => O.Tag.ToUpper().Equals("RG")).ToList().Any(a => a.Value.Equals(ReadGroup));
}
if (filter && !string.IsNullOrEmpty(Region))
{
if (alignedSequence.RName.Equals(region.Chromosome))
{
if (region.Start > -1)
{
if (alignedSequence.Pos >= region.Start)
{
if (region.End > -1)
{
if (alignedSequence.Pos <= region.End)
{
filter = true;
}
else
{
filter = false;
}
}
else
{
filter = true;
}
}
else
{
filter = false;
}
}
else
{
filter = true;
}
}
else
{
filter = false;
}
}
return filter;
}
/// <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>
public static int GetInsertLength(SAMAlignedSequence read1, SAMAlignedSequence read2)
{
return GetInsertLength(read1, read2, false);
}
/// <summary>
/// Gets an instance of SequenceRange class which represets alignment reigon of
/// specified aligned sequence (read) with reference sequence.
/// </summary>
/// <param name="alignedSequence">Aligned sequence.</param>
private static ISequenceRange GetRegion(SAMAlignedSequence alignedSequence)
{
string refSeqName = alignedSequence.RName;
long startPos = alignedSequence.Pos;
long endPos = alignedSequence.RefEndPos;
return new SequenceRange(refSeqName, startPos, endPos);
}
/// <summary>
/// Gets the paired reads type.
/// </summary>
/// <param name="read1">First aligned sequence.</param>
/// <param name="read2">Second aligned sequence.</param>
/// <param name="libraryInfo">Library information.</param>
public static PairedReadType GetPairedReadType(SAMAlignedSequence read1, SAMAlignedSequence read2, CloneLibraryInformation libraryInfo)
{
if (libraryInfo == null)
{
throw new ArgumentNullException("libraryInfo");
}
return GetPairedReadType(read1, read2, libraryInfo.MeanLengthOfInsert, libraryInfo.StandardDeviationOfInsert);
}
/// <summary>
/// Gets a boolean value which indicates that whether the specified read is reverse oriented or not.
/// </summary>
/// <param name="read">Aligned Sequence.</param>
public static bool IsReverseRead(SAMAlignedSequence read)
{
if (read == null)
{
throw new ArgumentNullException("read");
}
return !IsForwardRead(read);
}
/// <summary>
/// Gets a boolean value which indicates that whether the specified read is forward oriented or not.
/// </summary>
/// <param name="read">Aligned Sequence.</param>
public static bool IsForwardRead(SAMAlignedSequence read)
{
if (read == null)
{
throw new ArgumentNullException("read");
}
return (read.Flag & SAMFlags.QueryOnReverseStrand) == 0;
}
/// <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;
}
}
return read1.ISize == -read2.ISize ? (read1.ISize >= 0 ? read1.ISize : -read1.ISize) : 0;
}
// Gets the length of the optional fields in a SAMAlignedSequence object.
private static int GetAuxiliaryDataLength(SAMAlignedSequence alignedSeq)
{
int size = 0;
foreach (SAMOptionalField field in alignedSeq.OptionalFields)
{
size += 3;
int valueSize = GetOptionalFieldValueSize(field);
if (valueSize == 0)
{
string message = string.Format(CultureInfo.InvariantCulture, Properties.Resource.BAM_InvalidIntValueInOptFieldOfAlignedSeq, field.Value, field.Tag, alignedSeq.QName);
throw new FormatException(message);
}
size += valueSize < 0 ? -valueSize : valueSize;
}
return size;
}
/// <summary>
/// Add a sequence to the filtered output file header
/// </summary>
private void AddToHeader(SAMAlignedSequence seq)
{
newHeader.ReferenceSequences.Add(new ReferenceSequenceInfo(seq.RName, GetSequence(seq).Length));
// for each good cluster
SAMRecordField sq = new SAMRecordField("SQ");
sq.Tags.Add(new SAMRecordFieldTag("SN", seq.RName));
sq.Tags.Add(new SAMRecordFieldTag("LN", GetSequence(seq).Length.ToString(ci)));
newHeader.RecordFields.Add(sq);
}
/// <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);
}
/// <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;
}
public static void WriteFasta(this SAMAlignedSequence sam, StreamWriter sw)
{
sw.WriteLine(">" + sam.QName);
sw.WriteLine(sam.GetQuerySequenceString());
}
/// <summary>
/// General method to Invalidate Quality Sequences
/// <param name="method">enum type to execute different overload</param>
/// </summary>
private static void ValidateQualitySeqLength(ParseOrFormatQualLength method)
{
SAMAlignedSequence align = new SAMAlignedSequence();
try
{
switch (method)
{
case ParseOrFormatQualLength.AlignedSeq:
SAMParser.ParseQualityNSequence(
align,
Alphabets.DNA,
null,
String.Empty);
break;
case ParseOrFormatQualLength.Sequencedata:
align.QName = "Quality Value";
SAMParser.ParseQualityNSequence(
align,
Alphabets.DNA,
null,
String.Empty);
break;
case ParseOrFormatQualLength.Qualitydata:
align.QName = "Quality Value";
SAMParser.ParseQualityNSequence(
align,
Alphabets.DNA,
null,
Constants.QualitySequence);
break;
case ParseOrFormatQualLength.QualityLength:
align.QName = "Quality Value";
SAMParser.ParseQualityNSequence(
align,
Alphabets.DNA,
null,
Constants.QualitySequence);
break;
default:
break;
}
Assert.Fail();
}
catch (ArgumentException)
{
ApplicationLog.WriteLine(
"SAM Parser P2 : Successfully validated the exception");
}
catch (FormatException)
{
ApplicationLog.WriteLine(
"SAM Parser P2 : Successfully validated the exception");
}
}
/// <summary>
/// Update the linear index array based on an aligned read and its current coordinates
/// </summary>
/// <param name="alignedSeq"></param>
/// <param name="offset"></param>
internal void UpdateLinearArrayIndex(SAMAlignedSequence alignedSeq, FileOffset offset)
{
int pos = alignedSeq.Pos > 0 ? alignedSeq.Pos - 1 : 0;
int end = alignedSeq.RefEndPos > 0 ? alignedSeq.RefEndPos - 1 : 0;
pos = pos >> 14;
end = end >> 14;
if (end > largestBinSeen) {largestBinSeen = end;}
for (int i = pos; i <= end; i++)
{
var cur = offSetArray[i];
//TODO: Is second check necessary? Seems to always be true as we are doing things in order
if (cur.BothDataElements == 0 || cur > offset) {
offSetArray[i] = offset;
}
}
}
/// <summary>
/// Returns an aligned sequence by parses the BAM file.
/// </summary>
private SAMAlignedSequence GetAlignedSequence(int start, int end)
{
byte[] array = new byte[4];
ReadUnCompressedData(array, 0, 4);
int blockLen = Helper.GetInt32(array, 0);
byte[] alignmentBlock = new byte[blockLen];
ReadUnCompressedData(alignmentBlock, 0, blockLen);
SAMAlignedSequence alignedSeq = new SAMAlignedSequence();
int value;
UInt32 UnsignedValue;
// 0-4 bytes
int refSeqIndex = Helper.GetInt32(alignmentBlock, 0);
if (refSeqIndex == -1)
alignedSeq.RName = "*";
else
alignedSeq.RName = refSeqNames[refSeqIndex];
// 4-8 bytes
alignedSeq.Pos = Helper.GetInt32(alignmentBlock, 4) + 1;
// if there is no overlap no need to parse further.
// BAMPos > closedEnd
// => (alignedSeq.Pos - 1) > end -1
if (alignedSeq.Pos > end)
{
return null;
}
// 8 - 12 bytes "bin<<16|mapQual<<8|read_name_len"
UnsignedValue = Helper.GetUInt32(alignmentBlock, 8);
// 10 -12 bytes
alignedSeq.Bin = (int)(UnsignedValue & 0xFFFF0000) >> 16;
// 9th bytes
alignedSeq.MapQ = (int)(UnsignedValue & 0x0000FF00) >> 8;
// 8th bytes
int queryNameLen = (int)(UnsignedValue & 0x000000FF);
// 12 - 16 bytes
UnsignedValue = Helper.GetUInt32(alignmentBlock, 12);
// 14-16 bytes
int flagValue = (int)(UnsignedValue & 0xFFFF0000) >> 16;
alignedSeq.Flag = (SAMFlags)flagValue;
// 12-14 bytes
int cigarLen = (int)(UnsignedValue & 0x0000FFFF);
// 16-20 bytes
int readLen = Helper.GetInt32(alignmentBlock, 16);
// 20-24 bytes
int mateRefSeqIndex = Helper.GetInt32(alignmentBlock, 20);
if (mateRefSeqIndex != -1)
{
alignedSeq.MRNM = refSeqNames[mateRefSeqIndex];
}
else
{
alignedSeq.MRNM = "*";
}
// 24-28 bytes
alignedSeq.MPos = Helper.GetInt32(alignmentBlock, 24) + 1;
// 28-32 bytes
alignedSeq.ISize = Helper.GetInt32(alignmentBlock, 28);
// 32-(32+readLen) bytes
alignedSeq.QName = System.Text.ASCIIEncoding.ASCII.GetString(alignmentBlock, 32, queryNameLen - 1);
StringBuilder strbuilder = new StringBuilder();
int startIndex = 32 + queryNameLen;
for (int i = startIndex; i < (startIndex + cigarLen * 4); i += 4)
{
// Get the CIGAR operation length stored in first 28 bits.
UInt32 cigarValue = Helper.GetUInt32(alignmentBlock, i);
strbuilder.Append(((cigarValue & 0xFFFFFFF0) >> 4).ToString(CultureInfo.InvariantCulture));
// Get the CIGAR operation stored in last 4 bits.
value = (int)cigarValue & 0x0000000F;
// MIDNSHP=>0123456
switch (value)
{
case 0:
strbuilder.Append("M");
break;
case 1:
strbuilder.Append("I");
break;
case 2:
strbuilder.Append("D");
break;
case 3:
strbuilder.Append("N");
break;
case 4:
strbuilder.Append("S");
break;
case 5:
strbuilder.Append("H");
break;
case 6:
strbuilder.Append("P");
break;
case 7:
strbuilder.Append("=");
break;
case 8:
strbuilder.Append("X");
break;
default:
throw new FileFormatException(Properties.Resource.BAM_InvalidCIGAR);
}
}
string cigar = strbuilder.ToString();
if (string.IsNullOrWhiteSpace(cigar))
{
alignedSeq.CIGAR = "*";
}
else
{
alignedSeq.CIGAR = cigar;
}
// if there is no overlap no need to parse further.
// ZeroBasedRefEnd < start
// => (alignedSeq.RefEndPos -1) < start
if (alignedSeq.RefEndPos - 1 < start && alignedSeq.RName!=Properties.Resource.SAM_NO_REFERENCE_DEFINED_INDICATOR)
{
return null;
}
startIndex += cigarLen * 4;
strbuilder = new StringBuilder();
int index = startIndex;
for (; index < (startIndex + (readLen + 1) / 2) - 1; index++)
{
// Get first 4 bit value
value = (alignmentBlock[index] & 0xF0) >> 4;
strbuilder.Append(GetSeqChar(value));
// Get last 4 bit value
value = alignmentBlock[index] & 0x0F;
strbuilder.Append(GetSeqChar(value));
}
value = (alignmentBlock[index] & 0xF0) >> 4;
strbuilder.Append(GetSeqChar(value));
if (readLen % 2 == 0)
{
value = alignmentBlock[index] & 0x0F;
strbuilder.Append(GetSeqChar(value));
}
startIndex = index + 1;
string strSequence = strbuilder.ToString();
byte[] qualValues = new byte[readLen];
string strQualValues = "*";
if (alignmentBlock[startIndex] != 0xFF)
{
for (int i = startIndex; i < (startIndex + readLen); i++)
{
qualValues[i - startIndex] = (byte)(alignmentBlock[i] + 33);
}
strQualValues = System.Text.ASCIIEncoding.ASCII.GetString(qualValues);
}
SAMParser.ParseQualityNSequence(alignedSeq, Alphabet, strSequence, strQualValues);
startIndex += readLen;
if (alignmentBlock.Length > startIndex + 4 && alignmentBlock[startIndex] != 0x0 && alignmentBlock[startIndex + 1] != 0x0)
{
for (index = startIndex; index < alignmentBlock.Length; )
{
SAMOptionalField optionalField = new SAMOptionalField();
optionalField.Tag = System.Text.ASCIIEncoding.ASCII.GetString(alignmentBlock, index, 2);
index += 2;
char vType = (char)alignmentBlock[index++];
string valueType = vType.ToString();
// SAM format supports [AifZH] for value type.
// In BAM, an integer may be stored as a signed 8-bit integer (c), unsigned 8-bit integer (C), signed short (s), unsigned
// short (S), signed 32-bit (i) or unsigned 32-bit integer (I), depending on the signed magnitude of the integer. However,
// in SAM, all types of integers are presented as type ʻiʼ.
string message = Helper.IsValidPatternValue("VType", valueType, BAMOptionalFieldRegex);
if (!string.IsNullOrEmpty(message))
{
throw new FormatException(message);
}
optionalField.Value = GetOptionalValue(vType, alignmentBlock, ref index).ToString();
// Convert to SAM format.
if ("cCsSI".IndexOf(vType) >= 0)
{
valueType = "i";
}
optionalField.VType = valueType;
alignedSeq.OptionalFields.Add(optionalField);
}
}
return alignedSeq;
}
/// <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="sequencedata">Sequence data.</param>
/// <param name="qualitydata">Quality values.</param>
public static void ParseQualityNSequence(SAMAlignedSequence alignedSeq, IAlphabet alphabet, string sequencedata, string qualitydata)
{
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, Properties.Resource.FastQ_InvalidQualityScoresLength, alignedSeq.QName);
message = string.Format(CultureInfo.CurrentCulture, Properties.Resource.IOFormatErrorMessage, Properties.Resource.SAM_NAME, message1);
Trace.Report(message);
throw new FileFormatException(message);
}
}
ISequence sequence = null;
if (isQualitativeSequence)
{
QualitativeSequence qualSeq = new QualitativeSequence(alphabet, fastQType, sequencedata, ASCIIEncoding.ASCII.GetString(qualScores));
qualSeq.ID = alignedSeq.QName;
sequence = qualSeq;
}
else
{
sequence = new Sequence(alphabet, sequencedata);
sequence.ID = alignedSeq.QName;
}
alignedSeq.QuerySequence = sequence;
}
// Validates the alignment.
private SequenceAlignmentMap ValidateAlignment(ISequenceAlignment sequenceAlignment)
{
SequenceAlignmentMap seqAlignmentMap = sequenceAlignment as SequenceAlignmentMap;
if (seqAlignmentMap != null)
{
ValidateAlignmentHeader(seqAlignmentMap.Header);
if (CreateSortedBAMFile && SortType == BAMSortByFields.ChromosomeNameAndCoordinates)
{
this.refSequences = SortSequenceRanges(seqAlignmentMap.Header.GetReferenceSequenceRanges());
}
else
{
this.refSequences = seqAlignmentMap.Header.GetReferenceSequenceRanges();
}
return seqAlignmentMap;
}
SAMAlignmentHeader header = sequenceAlignment.Metadata[Helper.SAMAlignmentHeaderKey] as SAMAlignmentHeader;
if (header == null)
{
throw new ArgumentException(Properties.Resource.SAMAlignmentHeaderNotFound);
}
ValidateAlignmentHeader(header);
seqAlignmentMap = new SequenceAlignmentMap(header);
if (CreateSortedBAMFile && SortType == BAMSortByFields.ChromosomeNameAndCoordinates)
{
this.refSequences = SortSequenceRanges(seqAlignmentMap.Header.GetReferenceSequenceRanges());
}
else
{
this.refSequences = seqAlignmentMap.Header.GetReferenceSequenceRanges();
}
foreach (IAlignedSequence alignedSeq in sequenceAlignment.AlignedSequences)
{
SAMAlignedSequenceHeader alignedHeader = alignedSeq.Metadata[Helper.SAMAlignedSequenceHeaderKey] as SAMAlignedSequenceHeader;
if (alignedHeader == null)
{
throw new ArgumentException(Properties.Resource.SAMAlignedSequenceHeaderNotFound);
}
SAMAlignedSequence samAlignedSeq = new SAMAlignedSequence(alignedHeader);
samAlignedSeq.QuerySequence = alignedSeq.Sequences[0];
seqAlignmentMap.QuerySequences.Add(samAlignedSeq);
}
return seqAlignmentMap;
}
/// <summary>
/// Writes aligned sequence to output stream.
/// </summary>
/// <param name="header">Alignment header.</param>
/// <param name="alignedSequence">Aligned sequence to write.</param>
private void WriteAlignedSequence(SAMAlignmentHeader header, SAMAlignedSequence alignedSequence)
{
if (UnCompressedBAM || BAMOutput)
{
// In case of compressed bamoutput uncompressed file will be compressed before sending it to output stream.
bamformatter.WriteAlignedSequence(header, alignedSequence, bamUncompressedOutStream);
}
else
{
SAMFormatter.WriteSAMAlignedSequence(writer, alignedSequence);
}
}
/// <summary>
/// Writes SAMAlignedSequence to specified stream.
/// </summary>
/// <param name="header">Header from SAM object.</param>
/// <param name="alignedSeq">SAMAlignedSequence object.</param>
/// <param name="writer">Stream to write.</param>
public void WriteAlignedSequence(SAMAlignmentHeader header, SAMAlignedSequence alignedSeq, Stream writer)
{
if (header == null)
{
throw new ArgumentNullException("header");
}
if (alignedSeq == null)
{
throw new ArgumentNullException("alignedSeq");
}
if (writer == null)
{
throw new ArgumentNullException("writer");
}
if (this.refSequences == null)
{
this.refSequences = header.GetReferenceSequenceRanges();
}
WriteAlignedSequence(alignedSeq, writer);
}
/// <summary>
/// Method throws an exception if sequence violates any assumption made by this class anywhere.
/// Avoids, separate checks within each method.
/// </summary>
/// <param name="seq"></param>
private void validateSequence(SAMAlignedSequence seq)
{
if (seq == null) {
throw new ArgumentNullException("seq");
}
if (String.IsNullOrEmpty(seq.RName) ||
seq.RefEndPos <= seq.Pos ||
String.IsNullOrEmpty(seq.CIGAR) ||
seq.CIGAR =="*" ||
!(seq.QuerySequence is QualitativeSequence) )
{
throw new ArgumentException("Tried to build a pileup with an invalid sequence. Sequence was:\n"+
seq.ToString());
}
}
/// <summary>
/// Writes SAMAlignedSequence to specified stream.
/// </summary>
/// <param name="alignedSeq">SAMAlignedSequence object.</param>
/// <param name="writer">Stream to write.</param>
private void WriteAlignedSequence(SAMAlignedSequence alignedSeq, Stream writer)
{
// Get the total block size required.
int blocksize = GetBlockSize(alignedSeq);
// Get Reference sequence index.
int rid = GetRefSeqID(alignedSeq.RName);
// bin<<16|mapQual<<8|read_name_len (including NULL)
uint bin_mq_nl = (uint)alignedSeq.Bin << 16;
bin_mq_nl = bin_mq_nl | (uint)alignedSeq.MapQ << 8;
bin_mq_nl = bin_mq_nl | (uint)(alignedSeq.QName.Length + 1);
// flag<<16|cigar_len
uint flag_nc = (uint)alignedSeq.Flag << 16;
flag_nc = flag_nc | (uint)GetCIGARLength(alignedSeq.CIGAR);
int readLen = (int)alignedSeq.QuerySequence.Count;
int mateRefId = GetRefSeqID(alignedSeq.MRNM);
byte[] readName = Encoding.UTF8.GetBytes(alignedSeq.QName);
// Cigar: op_len<<4|op. Op: MIDNSHP=X => 012345678
IList<uint> encodedCIGAR = GetEncodedCIGAR(alignedSeq.CIGAR);
//block size
writer.Write(Helper.GetLittleEndianByteArray(blocksize), 0, 4);
// Reference sequence index.
writer.Write(Helper.GetLittleEndianByteArray(rid), 0, 4);
// Pos
writer.Write(Helper.GetLittleEndianByteArray(alignedSeq.Pos > 0 ? alignedSeq.Pos - 1 : -1), 0, 4);
// bin<<16|mapQual<<8|read_name_len (including NULL)
writer.Write(Helper.GetLittleEndianByteArray(bin_mq_nl), 0, 4);
// flag<<16|cigar_len
writer.Write(Helper.GetLittleEndianByteArray(flag_nc), 0, 4);
// Length of the read
writer.Write(Helper.GetLittleEndianByteArray(readLen), 0, 4);
// Mate reference sequence index
writer.Write(Helper.GetLittleEndianByteArray(mateRefId), 0, 4);
// mate_pos - Leftmost coordinate of the mate
// As per SAM format Mpos will be 1 based and 0 indicates unpaired or pairing information is unavailabe.
// In case of BAM format Mpos will be zero based and -1 indicates unpaired or pairing information is unavailabe.
writer.Write(Helper.GetLittleEndianByteArray(alignedSeq.MPos - 1), 0, 4);
// Insert size of the read pair (if paired)
writer.Write(Helper.GetLittleEndianByteArray(alignedSeq.ISize), 0, 4);
// Read name, null terminated
writer.Write(readName, 0, readName.Length);
writer.WriteByte((byte)'\0');
// Cigar: op_len<<4|op. Op: MIDNSHP=>0123456
foreach (uint data in encodedCIGAR)
{
writer.Write(Helper.GetLittleEndianByteArray(data), 0, 4);
}
// 4-bit encoded read: =ACGTN=>0,1,2,4,8,15; the earlier base is stored in the high-order 4 bits of the byte.
byte[] encodedValues = GetEncodedSequence(alignedSeq);
writer.Write(encodedValues, 0, encodedValues.Length);
// Phred base quality (0xFF if absent)
encodedValues = GetQualityValue(alignedSeq.QuerySequence);
writer.Write(encodedValues, 0, encodedValues.Length);
// Optional fields
foreach (SAMOptionalField field in alignedSeq.OptionalFields)
{
byte[] optionalArray = GetOptioanField(field);
writer.Write(optionalArray, 0, optionalArray.Length);
}
}
/// <summary>
/// Add a sequence into a dictionary value item which represents a list of sequences. The list to which to add
/// to is found using key
/// </summary>
private static void AddToDict(Dictionary<String, List<SAMAlignedSequence>> dict, string key, SAMAlignedSequence seq)
{
if (key != null)
{
if (dict.ContainsKey(key))
{
List<SAMAlignedSequence> existingVal = dict[key];
existingVal.Add(seq);
dict[key] = existingVal;
}
else
{
dict.Add(key, new List<SAMAlignedSequence> { seq });
}
}
else
{
throw new ArgumentException(Properties.Resources.INVALID_KEY);
}
}
/// <summary>
/// Add a sequence. If the sequence belongs to the current cluster, store it. If the sequence
/// is part of a new cluster, process the current sequence cluster then add the sequence
/// to a new cluster
/// </summary>
/// <param name="sequence">A sequence.</param>
/// <returns>Returns true if the sequence could be added, false if the handler has been closed.</returns>
public bool Add(SAMAlignedSequence sequence)
{
if (allSequences == null)
{
allSequences = new Collection<SAMAlignedSequence>();
}
if(sequence == null)
{
return true;
}
if (!finished)
{
string thisSeqCluster = sequence.RName; // Cluster the sequence we just added belongs to
// This is the first sequence for the first cluster
if (currentClusterId == null)
{
currentClusterId = thisSeqCluster;
}
// This sequence belongs to a different cluster from the ones currently stored by this handler
// (Process currently stored sequences before adding the new sequence)
else if (!currentClusterId.Equals(thisSeqCluster))
{
++numberClustersParsed; // mark off another cluster
ProcessSequences();
currentClusterId = thisSeqCluster;
allSequences = new Collection<SAMAlignedSequence>();
}
allSequences.Add(sequence);
return true;
}
// Processing of sequences should be finished but we are still outputting to the bam file
// Or we are supposed to write to a bam file and the header and body files have not yet been merged
// Wait for output to the bam file to complete
else if ((!canWriteToBam && writeToFilteredBam) || (!bamFilesMerged && writeToFilteredBam))
{
while (!canWriteToBam || !bamFilesMerged)
{
Thread.Sleep(20000); // sleep 20 seconds
}
return true;
}
// finished == true, bam file is writable and bam files have been merged (or no bam file was ever written to)
// returning false indicates to calling process that no more sequences will be accepted
else
{
return false;
}
}
/// <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>
/// Given a sequence, returns the reference ID, or null if sequences are unmapped
/// </summary>
private static string GetId(SAMAlignedSequence sequence)
{
if (!sequence.Flag.HasFlag(SAMFlags.UnmappedQuery))
{
return (sequence != null) ? sequence.RName : null;
}
else
{
return null;
}
}
/// <summary>
/// Initializes a new instance of the <see cref="Bio.IO.PacBio.PacBioCCSRead"/> class. From an initially parsed BAM file.
/// </summary>
/// <param name="s">S.</param>
public PacBioCCSRead (SAMAlignedSequence s)
{
/* TODO: Converting from binary to string and back is beyond silly...
* no performance hit worth worrying about at present, but in the future it might be worth
* going directly from binary to the type rather than through string intermediates */
foreach (var v in s.OptionalFields) {
if (v.Tag == "sn") {
var snrs = v.Value.Split (',').Skip (1).Select (x => Convert.ToSingle (x)).ToArray ();
SnrA = snrs [0];
SnrC = snrs [1];
SnrG = snrs [2];
SnrT = snrs [3];
} else if (v.Tag == "zm") {
HoleNumber = (int)Convert.ToInt32 (v.Value);
} else if (v.Tag == "pq") {
// This tag is now deprecated by the rq tag
ReadQuality = Convert.ToSingle (v.Value);
} else if (v.Tag == "rq") {
ReadQuality = Convert.ToSingle (v.Value);
}else if (v.Tag == "za") {
AvgZscore = (float)Convert.ToSingle (v.Value);
} else if (v.Tag == "rs") {
statusCounts = v.Value.Split (',').Skip (1).Select (x => Convert.ToInt32 (x)).ToArray ();
} else if (v.Tag == "np") {
NumPasses = Convert.ToInt32 (v.Value);
} else if (v.Tag == "RG") {
ReadGroup = v.Value;
} else if (v.Tag == "zs") {
ZScores = v.Value.Split (',').Skip (1).Select (x => Convert.ToSingle (x)).ToArray ();
}
}
// TODO: We should use String.Intern here, but not available in PCL...
// Movie = String.Intern(s.QuerySequence.ID.Split ('/') [0]);
Movie = s.QuerySequence.ID.Split ('/') [0];
Sequence = s.QuerySequence as QualitativeSequence;
}
/// <summary>
/// Given a SAMAlignedSequence, get the RG tag for that read
/// </summary>
private static string GetRgTag(SAMAlignedSequence seq)
{
foreach (SAMOptionalField field in seq.OptionalFields)
{
// I iterate through to find RG each time in case the optional fields
// do not have a consistent format.
if (field.Tag == "RG")
{
return field.Value;
}
}
return null;
}
/// <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>
/// Given a SAMAlignedSequence, returns a string representation of the genetic sequence
/// </summary>
private static string GetSequence(SAMAlignedSequence seq)
{
String seqStr = seq.QuerySequence.ToString();
return Regex.Split(seqStr, "\r\n")[0];
}