/// <summary>
/// Converts the input BAM to SAM file format.
/// </summary>
private void ConvertFromBAMToSAM()
{
using (Stream stream = new FileStream(InputFilePath, FileMode.Open, FileAccess.Read))
{
SAMAlignmentHeader header = null;
try
{
header = bamparser.GetHeader(stream);
}
catch (Exception ex)
{
throw new InvalidOperationException(Resources.InvalidBAMFile, ex);
}
WriteHeader(header);
if (!HeaderOnly)
{
if (!string.IsNullOrEmpty(Library))
{
rgRecFields = header.RecordFields.Where(R => R.Typecode.ToUpper().Equals("RG")).ToList();
}
foreach (SAMAlignedSequence alignedSequence in GetAlignedSequence(stream))
{
WriteAlignedSequence(header, alignedSequence);
}
}
}
}
/// <summary>
/// Read input file header and set progress bar based on number of sequences in input file
/// </summary>
private void ReadHeader(string filename, BAMParser parser, int numClustersInInputFile)
{
using (Stream readStream = new FileStream(filename, FileMode.Open, FileAccess.Read, FileShare.Read))
{
metricHandler.InputHeader = parser.GetHeader(readStream);
numClustersInInputFile = metricHandler.InputHeader.ReferenceSequences.Count;
Console.WriteLine(Properties.Resources.CLUSTER_COUNT_DISPLAY + numClustersInInputFile);
}
Dispatcher.BeginInvoke(System.Windows.Threading.DispatcherPriority.Normal,
new IntDelegate(SetProgressBar), numClustersInInputFile);
}
public List <string> ReadHeaders()
{
var header = _parser.GetHeader(_stream);
var rfs = (from h in header.RecordFields
select string.Format("@{0}\t{1}", h.Typecode, (from t in h.Tags
select string.Format("{0}:{1}", t.Tag, t.Value)).Merge("\t"))).ToList();
var sqs = (from s in header.ReferenceSequences
select string.Format("@SQ\tSN:{0}\tLN:{1}", s.Name, s.Length)).ToList();
return(rfs.Union(sqs).Union(header.Comments).ToList());
}
void validateInputFileAndLoadSampleNames()
{
if (!File.Exists(InputFilename))
{
throw new FileNotFoundException("Could not find file: " + InputFilename);
}
using (Stream stream = new FileStream(InputFilename, FileMode.Open, FileAccess.Read))
{
BAMParser bp = new BAMParser();
header = bp.GetHeader(stream);
var tmp = header.RecordFields.Where(x => x.Typecode == "RG").ToList();
sampleNames = tmp.Select(z => z.Tags.Where(p => p.Tag == "ID").First()).Select(z => z.Value).ToList();
NumberOfReadGroups = sampleNames.Count;
NumerOfSequences = header.ReferenceSequences.Count;
}
Console.WriteLine("Processing file with " + sampleNames.Count.ToString() + " samples and " + NumerOfSequences.ToString() + " reference sequences.");
}
/// <summary>
/// Parse the CCS reads in a PacBio CCS BAM File.
/// </summary>
/// <returns>The CCS reads.</returns>
/// <param name="stream">A stream to parse.</param>
public static IEnumerable <PacBioCCSRead> Parse(Stream stream)
{
if (stream == null)
{
throw new ArgumentNullException("stream");
}
BAMParser bp = new BAMParser();
var header = bp.GetHeader(stream);
var field = header.RecordFields.ToList();
var pg = field.Where(p => p.Typecode == "PG").FirstOrDefault();
if (pg == null)
{
throw new ArgumentException("BAM file did not contain a 'PG' tag in header");
}
var cl = pg.Tags.Where(z => z.Tag == "CL").FirstOrDefault();
if (cl == null)
{
throw new ArgumentException("BAM file did not contain a 'CL' tag within the 'PG' group in header.");
}
var cmd = cl.Value;
if (!cmd.StartsWith("ccs"))
{
throw new ArgumentException("This is not a BAM file produced by the ccs command.");
}
stream.Seek(0, SeekOrigin.Begin);
foreach (var s in bp.Parse(stream))
{
yield return(new PacBioCCSRead(s as SAMAlignedSequence));
}
}
/// <summary>
/// Extract/print all or sub alignments in SAM or BAM format.
/// By default, this command assumes the file on the command line is in
/// BAM format and it prints the alignments in SAM.
/// SAMUtil.exe view in.bam
/// </summary>
public void ViewResult()
{
try
{
if (string.IsNullOrEmpty(InputFilePath))
{
throw new InvalidOperationException("Input File Not specified");
}
if (!string.IsNullOrEmpty(Region))
{
StringToRegionConverter();
}
Initialize();
SAMAlignmentHeader header = null;
if (!SAMInput)
{
Stream stream = new FileStream(InputFilePath, FileMode.Open, FileAccess.Read);
try
{
header = bamparser.GetHeader(stream);
}
catch
{
throw new InvalidOperationException(Resources.InvalidBAMFile);
}
WriteHeader(header);
if (!HeaderOnly)
{
if (!string.IsNullOrEmpty(Library))
{
rgRecFields = header.RecordFields.Where(R => R.Typecode.ToUpper().Equals("RG")).ToList();
}
foreach (SAMAlignedSequence alignedSequence in GetAlignedSequence(stream))
{
WriteAlignedSequence(header, alignedSequence);
}
}
}
else
{
try
{
header = SAMParser.ParseSAMHeader(InputFilePath);
}
catch
{
throw new InvalidOperationException(Resources.InvalidSAMFile);
}
if (header == null)
{
throw new InvalidOperationException("SAM file doesn't contian header");
}
WriteHeader(header);
if (!HeaderOnly)
{
if (!string.IsNullOrEmpty(Library))
{
rgRecFields = header.RecordFields.Where(R => R.Typecode.ToUpper().Equals("RG")).ToList();
}
using (StreamReader textReader = new StreamReader(InputFilePath))
{
foreach (SAMAlignedSequence alignedSeq in GetAlignedSequence(textReader))
{
WriteAlignedSequence(header, alignedSeq);
}
}
}
if (UnCompressedBAM)
{
bamUncompressedOutStream.Flush();
if (writer != null)
{
DisplayBAMContent(bamUncompressedOutStream);
}
}
if (BAMOutput && !UnCompressedBAM)
{
bamUncompressedOutStream.Flush();
bamUncompressedOutStream.Seek(0, SeekOrigin.Begin);
bamformatter.CompressBAMFile(bamUncompressedOutStream, bamCompressedOutStream);
bamCompressedOutStream.Flush();
if (writer != null)
{
DisplayBAMContent(bamCompressedOutStream);
}
}
}
}
finally
{
Close();
}
}
/// <summary>
/// Run ConPADE on each contig of the input BAM file.
/// </summary>
/// <param name="bamName">Name of the input BAM file.</param>
public void RunFile(string bamName)
{
// Current implementation requires that minimum ploidy be 1
int min_ploidy = 1;
int number_of_ploidies = max_ploidy - min_ploidy + 1;
// Set nucleotide proportions (genotypes)
double[][][] nuc_props = Nuc_Props(min_ploidy, number_of_ploidies);
// Set dosage probabilities
double SNP_density = (double)1 / snpDens;
double no_SNP_prob = Math.Log((1 - SNP_density) / 2);
double[][] dose_probs = Dose_Probs(min_ploidy, number_of_ploidies, SNP_density, no_SNP_prob);
// Set HiSeq error model
double[, , , ,] log_probs = Error_Probs();
// Set substitution model
double[, ,] log_subst_probs = Subst_Probs();
// Set SNP calling probability
double log_SNP_thres = SNPthres * Math.Log(10) / -10;
Stopwatch clock = new Stopwatch();
Console.WriteLine("Program started at {0}\n", DateTime.Now);
Stream bam_stream = new FileStream(bamName, FileMode.Open, FileAccess.Read);
BAMParser parser = new BAMParser();
SAMAlignmentHeader header = parser.GetHeader(bam_stream);
string temp = Path.GetFileNameWithoutExtension(bamName);
// Find first valid alignment in BAM file
SAMAlignedSequence next_alignment = parser.GetAlignedSequence(true);
while (next_alignment == null || next_alignment.RName == "*" || next_alignment.IsDummyRead)
{
next_alignment = parser.GetAlignedSequence(true);
}
TextWriter writer_log_like = null;
TextWriter writer_SNP = null;
TextWriter writer_ploidy = null;
TextWriter writer_reads = null;
// Create global output files and write headers.
if (!splitContigs)
{
string SNP_file = temp + "_SNP.txt";
writer_SNP = new StreamWriter(SNP_file);
writer_SNP.WriteLine("Contig\tPosition\tAlleles\tCounts\tDosage\tPhredQuality");
string ploidy_file = temp + "_ploidy.txt";
writer_ploidy = new StreamWriter(ploidy_file);
writer_ploidy.Write("Contig\tBestPloidy");
for (int i = 0; i < number_of_ploidies; i++)
{
writer_ploidy.Write("\tlogLike_M{0}", i + min_ploidy);
}
writer_ploidy.WriteLine("");
string reads_file = temp + "_readStats.txt";
writer_reads = new StreamWriter(reads_file);
writer_reads.WriteLine("Contig\tAlignedReads\tAlignedBases\tUsedReads\tUsedBases");
}
// Run over each contig in input BAM file.
int contig_ind = -1;
while (next_alignment != null && next_alignment.RName != "*" && !next_alignment.IsDummyRead)
{
string contig_name = next_alignment.RName;
Console.WriteLine("Started contig {0} at {1}",
contig_name, DateTime.Now);
clock.Restart();
#region Variables and file handles for current contig
long number_of_aligned_reads = 0;
long number_of_aligned_base_pairs = 0;
long number_of_used_reads = 0;
long number_of_used_base_pairs = 0;
// Create individual output files for the current contig.
if (splitContigs)
{
string name = temp + "_" + contig_name;
string log_like_file = name + "_log_likelihoods.txt";
writer_log_like = new StreamWriter(log_like_file);
string SNP_file = name + "_SNP.txt";
writer_SNP = new StreamWriter(SNP_file);
string ploidy_file = name + "_ploidy.txt";
writer_ploidy = new StreamWriter(ploidy_file);
string reads_file = name + "_readStats.txt";
writer_reads = new StreamWriter(reads_file);
}
double[] global_log_like = new double[number_of_ploidies];
while (header.ReferenceSequences[++contig_ind].Name != contig_name)
{
;
}
long contig_length = header.ReferenceSequences[contig_ind].Length;
// Create a queue to include all reads that overlap with a given position.
Queue <Padded_Read> read_queue = new Queue <Padded_Read>();
// Create a queue to include best doses for each tested position.
Queue <Best_Dose> dose_queue = new Queue <Best_Dose>((int)contig_length);
#endregion Variables and file handles for current contig
int positions_to_compute = 0;
long current_position = 0;
#region Run over every position in contig
while (current_position < contig_length)
{
if ((current_position % 1000000) == 0 && current_position != 0)
{
Console.WriteLine("At position {0} of {1}", current_position + 1, contig_length);
}
// Search for reads starting at current position.
Search_Reads(parser, ref next_alignment, contig_name, ref number_of_aligned_reads,
ref number_of_aligned_base_pairs, ref number_of_used_reads, ref number_of_used_base_pairs,
read_queue, current_position);
if (read_queue.Count > 0)
{
positions_to_compute++;
// Extract information from each read in queue.
byte[] obs_nucs;
byte[] is_GG;
bool[] reverse;
int[] quality_scores;
int[] neigh_quality_scores;
int[] scores;
int[] counts;
int k;
Extract_Read_Info(read_queue, current_position, out obs_nucs, out is_GG, out reverse,
out quality_scores, out neigh_quality_scores, out scores, out counts, out k);
// Find two most abundant nucleotides for this position.
byte nuc_one;
byte nuc_two;
Get_Two_Nucs(scores, out nuc_one, out nuc_two);
// Calculate Pr(obs|allele1) and Pr(obs|allele2).
double[][] log_nuc_probs = Obs_Probs(log_probs, log_subst_probs, obs_nucs, is_GG, reverse,
quality_scores, neigh_quality_scores, counts, k, nuc_one, nuc_two);
// Calculate log_likelihoods of genotypes for current position.
double[][] log_likelihoods = Log_Likelihoods(min_ploidy, max_ploidy, log_nuc_probs, nuc_props);
// Calculate log_likelihood of each ploidy and keep most likely allele dosage.
Global_Likelihood_Keep_Dose(min_ploidy, number_of_ploidies, dose_probs, global_log_like,
dose_queue, current_position, counts, nuc_one, nuc_two, log_likelihoods);
}
// Remove finished reads from queue. Finished reads no longer overlap with current position.
Padded_Read read_to_remove;
if (read_queue.Count > 0)
{
read_to_remove = read_queue.First();
}
else
{
read_to_remove = null;
}
while (read_to_remove != null &&
(read_to_remove.alignment.Pos + read_to_remove.alignment_length - 2) < current_position)
{
read_queue.Dequeue();
if (read_queue.Count > 0)
{
read_to_remove = read_queue.First();
}
else
{
read_to_remove = null;
}
}
++current_position;
}
#endregion Run over every position in contig
// Output log_likelihoods.
int best_log_like = 0;
for (int i = 0; i < number_of_ploidies; i++)
{
if (global_log_like[i] > global_log_like[best_log_like])
{
best_log_like = i;
}
if (splitContigs)
{
writer_log_like.WriteLine("Ploidy {0} - log_likelihood {1}", i + min_ploidy, global_log_like[i]);
}
}
// Output most likely ploidy.
int best_ploidy = best_log_like + min_ploidy;
if (splitContigs)
{
writer_ploidy.WriteLine(best_ploidy);
}
else
{
writer_ploidy.Write("{0}\t{1}", contig_name, best_ploidy);
for (int i = 0; i < number_of_ploidies; i++)
{
writer_ploidy.Write("\t{0}", global_log_like[i]);
}
writer_ploidy.WriteLine("");
}
// Output SNPs.
if (splitContigs)
{
writer_SNP.WriteLine("Position\tAlleles\tCounts\tDosage\tPhredQuality");
}
char[] nuc_chars = new char[4] {
'A', 'C', 'G', 'T'
};
foreach (Best_Dose cur_doses in dose_queue)
{
double cur_SNP_posterior = cur_doses.SNP_posterior[best_log_like];
if (cur_SNP_posterior <= log_SNP_thres)
{
int cur_best_dose = cur_doses.best_dose[best_log_like];
if (cur_best_dose != best_ploidy && cur_best_dose != 0)
{
if (splitContigs)
{
writer_SNP.WriteLine("{0}\t{1}|{2}\t{3}|{4}\t{5}\t{6}", cur_doses.position + 1,
nuc_chars[cur_doses.nuc_one], nuc_chars[cur_doses.nuc_two], cur_doses.count_one,
cur_doses.count_two, cur_best_dose, -10 * cur_SNP_posterior / Math.Log(10));
}
else
{
writer_SNP.WriteLine("{0}\t{1}\t{2}|{3}\t{4}|{5}\t{6}\t{7}", contig_name,
cur_doses.position + 1, nuc_chars[cur_doses.nuc_one], nuc_chars[cur_doses.nuc_two],
cur_doses.count_one, cur_doses.count_two, cur_best_dose,
-10 * cur_SNP_posterior / Math.Log(10));
}
}
}
}
// Output read statistics.
if (splitContigs)
{
writer_reads.WriteLine("\nNumber of aligned reads: {0}", number_of_aligned_reads);
writer_reads.WriteLine("Number of aligned base pairs: {0}", number_of_aligned_base_pairs);
writer_reads.WriteLine("\nNumber of used reads: {0}", number_of_used_reads);
writer_reads.WriteLine("Number of used base pairs: {0}", number_of_used_base_pairs);
}
else
{
writer_reads.WriteLine("{0}\t{1}\t{2}\t{3}\t{4}", contig_name, number_of_aligned_reads,
number_of_aligned_base_pairs, number_of_used_reads, number_of_used_base_pairs);
}
if (splitContigs)
{
writer_log_like.Close();
writer_SNP.Close();
writer_ploidy.Close();
writer_reads.Close();
}
clock.Stop();
Console.WriteLine("Time to run contig: {0} s\n", (double)clock.ElapsedMilliseconds / 1000);
}
if (!splitContigs)
{
writer_SNP.Close();
writer_ploidy.Close();
writer_reads.Close();
}
parser.Dispose();
Console.WriteLine("Finished at {0}\n", DateTime.Now);
}