static Dictionary <string, int> highRepSeqs; // long seqs + counts for high depth reads
static void Main(string[] args)
{
if (args.Length == 0)
{
Console.WriteLine("FilterReadsByDepth -min minDepth -max maxDepth [-reduce maxCopies] [-histoOnly] [-stats statsFN] [-f format] [-t #threads] cbtFN fileNames");
return;
}
List <string> FNParams = new List <string>(); // the .cbt name and the set of file names or patterns
int noThreads = 2;
// find out who we are so we can track what program & args produced the result files
Process myProcess = Process.GetCurrentProcess();
myProcessNameAndArgs = myProcess.ProcessName;
foreach (string a in args)
{
myProcessNameAndArgs = myProcessNameAndArgs + " " + a;
}
for (int p = 0; p < args.Length; p++)
{
if (args[p][0] == '-')
{
args[p] = args[p].ToLower();
if (args[p] == "-s" || args[p] == "-stats")
{
if (!CheckForParamValue(p, args.Length, "stats file name string expected after -s|-stats"))
{
return;
}
statsFN = args[p + 1];
p++;
continue;
}
if (args[p] == "-min")
{
if (!CheckForParamValue(p, args.Length, "minDepth number expected after -min"))
{
return;
}
try
{
minDepth = Convert.ToInt32(args[p + 1]);
}
catch
{
Console.WriteLine("expected a number for the -min parameter: " + args[p + 1]);
return;
}
p++;
continue;
}
if (args[p] == "-max")
{
if (!CheckForParamValue(p, args.Length, "maxDepth number expected after -max"))
{
return;
}
try
{
maxDepth = Convert.ToInt32(args[p + 1]);
}
catch
{
Console.WriteLine("expected a number for the -max parameter: " + args[p + 1]);
return;
}
p++;
continue;
}
if (args[p] == "-r" || args[p] == "-reduce")
{
if (!CheckForParamValue(p, args.Length, "reduced depth number expected after -reduce"))
{
return;
}
try
{
reducedDepth = Convert.ToInt32(args[p + 1]);
reducingReads = true;
}
catch
{
Console.WriteLine("expected a number for the -reduce parameter: " + args[p + 1]);
return;
}
p++;
continue;
}
if (args[p] == "-histoonly" || args[p] == "-ho")
{
histoOnly = true;
continue;
}
if (args[p] == "-t" || args[p] == "-threads")
{
if (!CheckForParamValue(p, args.Length, "number expected after -t|-threads"))
{
return;
}
try
{
noThreads = Convert.ToInt32(args[p + 1]);
}
catch
{
Console.WriteLine("expected a number for the -t|-threads parameter: " + args[p + 1]);
return;
}
p++;
continue;
}
if (args[p] == "-f" || args[p] == "-format")
{
if (!CheckForParamValue(p, args.Length, "reads format expected after -f|-format"))
{
return;
}
string readsFormatParam = args[p + 1].ToLower();
if (readsFormatParam == "fna")
{
readsFormat = MerStrings.formatFNA;
}
else if (readsFormatParam == "fasta")
{
readsFormat = MerStrings.formatFNA;
}
else if (readsFormatParam == "fa")
{
readsFormat = MerStrings.formatFNA;
}
else if (readsFormatParam == "fastq")
{
readsFormat = MerStrings.formatFASTQ;
}
else if (readsFormatParam == "fq")
{
readsFormat = MerStrings.formatFASTQ;
}
else
{
Console.WriteLine("reads format must be fasta or fastq: " + args[p + 1]);
return;
}
p++;
continue;
}
if (args[p] == "-o" || args[p] == "-output")
{
if (!CheckForParamValue(p, args.Length, "directory name expected after -o|-output"))
{
return;
}
outputDir = args[p + 1];
p++;
continue;
}
}
FNParams.Add(args[p]);
}
if (FNParams.Count < 2)
{
Console.WriteLine("expected a cbt file name and at least one reads file name or pattern");
return;
}
// validate the output directory & set the output prefix string
string fnSeparator = Path.DirectorySeparatorChar.ToString(); // \ for Windows; / for Unix/Linux
if (outputDir != null)
{
try
{
// add a trailing \ if the output directory name doesn't already have one
if (!outputDir.EndsWith(fnSeparator))
{
outputDir += fnSeparator;
}
string testOutputFN = outputDir + "43EDD23F-5F68-47f0-B7B9-66AE9EE3BF0B.txt";
StreamWriter testTemp = new StreamWriter(testOutputFN);
testTemp.Close();
File.Delete(testOutputFN);
}
catch
{
Console.WriteLine("Output directory: " + args[6] + " was invalid");
return;
}
}
// take the cbt file name from the start of the non-option list
string cbtFN = FNParams[0];
FNParams.RemoveAt(0);
if (FNParams.Count == 0)
{
Console.WriteLine("did not find any reads file names or patterns");
return;
}
if (!File.Exists(cbtFN))
{
Console.WriteLine("k-mer consensus (.cbt) file not found: " + cbtFN);
return;
}
List <string> readsFileNames = new List <string>(FNParams.Count);
List <string> readsFilePaths = new List <string>(FNParams.Count);
foreach (string readsFNP in FNParams)
{
string readsFileName;
string readsFilePath;
GetPathFN(readsFNP, out readsFilePath, out readsFileName);
readsFilePaths.Add(readsFilePath);
readsFileNames.Add(readsFileName);
}
List <string> expandedReadsFNs = new List <string>();
for (int f = 0; f < FNParams.Count; f++)
{
string[] matchedReadsFNs = Directory.GetFiles(readsFilePaths[f], readsFileNames[f], SearchOption.TopDirectoryOnly);
foreach (string matchedReadsFN in matchedReadsFNs)
{
expandedReadsFNs.Add(matchedReadsFN);
}
}
// make sure there aren't any duplicates in the file list (seems to be a bug on the Cherax SGI HPC system and it returns each file name twice)
List <string> distinctReadsFNs = new List <string>();
foreach (string fn in expandedReadsFNs)
{
if (!distinctReadsFNs.Contains(fn))
{
distinctReadsFNs.Add(fn);
}
}
// finally... the set of fully qualified, distinct reads files
string[] readsFNs;
readsFNs = distinctReadsFNs.ToArray();
Array.Sort(readsFNs);
int noOfReadsFiles = distinctReadsFNs.Count;
if (noOfReadsFiles == 0)
{
Console.WriteLine("No matching reads files found");
return;
}
StreamReader formatTester = new StreamReader(readsFNs[0]);
string firstLine = formatTester.ReadLine();
if (firstLine[0] == '>')
{
readsFormat = MerStrings.formatFNA;
}
if (firstLine[0] == '@')
{
readsFormat = MerStrings.formatFASTQ;
}
formatTester.Close();
formatTester = null;
if (statsFN == null)
{
// construct a stats
statsFN = readsFileNames[0].Substring(0, readsFileNames[0].LastIndexOf('.'));
statsFN = statsFN.Replace('?', '_');
statsFN = statsFN.Replace('*', '_');
statsFN = statsFN.Replace('/', '_');
statsFN = statsFN.Replace('\\', '_');
statsFN = statsFN.Replace("__", "_");
statsFN = statsFN.Replace("__", "_");
statsFN = statsFN + "_fstats.txt";
statsFN = statsFN.Replace("__", "_");
}
// calculate the min load depth from the min reps depth - don't need to load all of the singletons and other errors into memory
//int minLoadDepth = minDepth / 2;
//if (minLoadDepth <= 1)
// minLoadDepth = 2;
int minLoadDepth = minDepth;
long loadedUniqueMers = 0;
long loadedTotalMers = 0;
// load the .cbt file into a merTable (either a hash table (small) or a sorted array (large))
long mersLoaded = MerStrings.LoadCBTFile(cbtFN, minLoadDepth, 0, 0, minDepth,
out uniqueMers, out merSize, out averageDepth, out loadedUniqueMers, out loadedTotalMers);
if (merSize < 1 || merSize > 32)
{
Console.WriteLine("bad k-mer size found at start of .cbt file");
return;
}
MerStrings.Initialise(merSize);
highRepSeqs = new Dictionary <string, int>(10000000);
highRepSeqs.Add(new string('A', 40), 0);
highRepSeqs.Add(new string('C', 40), 0);
highRepSeqs.Add(new string('G', 40), 0);
highRepSeqs.Add(new string('T', 40), 0);
// resolve the FASTQ qual ambiguity by reading through quals until one is encountered that can only come from either of the alternative sets
if (readsFormat == MerStrings.formatFASTQ)
{
qualBase = MerStrings.ResolveFastqQualAmbiguity(readsFNs[0], out fullQualHeaders);
}
// and check whether we've got Unix data so we can write out the corrected files in the same format
string lfConvention = MerStrings.LFConvention(readsFNs[0]);
// start the monitor/synchronising thread
Thread monitorProgress = new Thread(RateReporter);
monitorProgress.Priority = ThreadPriority.AboveNormal;
monitorProgress.Start();
readsFiles = new StreamReader[2];
filteredReads = new StreamWriter[2];
Dictionary <int, int> readDepths = new Dictionary <int, int>(1000);
// filter a pair of files at a time (allowing us to filter many files in a single run while keeping pairedness)
for (int f = 0; f < noOfReadsFiles; f += 2)
{
// for each file in the pair
for (int p = 0; p < 2; p++)
{
if (f + p < noOfReadsFiles)
{
string fullReadsFN = readsFNs[f + p];
string readsPath;
string readsFN;
GetPathFN(fullReadsFN, out readsPath, out readsFN);
string fileSuffix = readsFN.Substring(readsFN.LastIndexOf('.'));
string fileWithoutSuffix = readsFN.Substring(0, readsFN.LastIndexOf('.'));
readsFiles[p] = new StreamReader(fullReadsFN, Encoding.ASCII, false, 1000000);
Console.WriteLine("filtering " + readsFN);
// check that the file appears to be in the expected format
char firstChar = (char)readsFiles[p].Peek();
if (readsFormat == MerStrings.formatFASTQ && firstChar != '@')
{
Console.WriteLine(readsFN + " does not appear to be in FASTQ format");
return;
}
if (readsFormat == MerStrings.formatFNA && firstChar != '>')
{
Console.WriteLine(readsFN + " does not appear to be in FASTA format");
return;
}
string outputPath = outputDir == null ? readsPath + fnSeparator : outputDir;
if (!histoOnly)
{
string maxDepthString = maxDepth.ToString();
if (maxDepth == int.MaxValue)
{
maxDepthString = "max";
}
filteredReads[p] = new StreamWriter(outputPath + fileWithoutSuffix + "_" + minDepth + "_" + maxDepthString + fileSuffix,
false, readsFiles[p].CurrentEncoding, 1000000);
filteredReads[p].NewLine = lfConvention;
}
}
else
{
readsFiles[p] = null;
filteredReads[p] = null;
}
}
filterThreadParams[] filterParams = new filterThreadParams[noThreads];
Thread[] filteringThreads = new Thread[noThreads];
// ready a new thread for each parallel healer
for (int b = 0; b < noThreads; b++)
{
filterParams[b] = new filterThreadParams();
filterParams[b].threadNumber = b + 1;
filterParams[b].readsFiles = readsFiles;
filterParams[b].filteredReads = filteredReads;
filteringThreads[b] = new Thread(new ParameterizedThreadStart(Program.FilteringThread));
filteringThreads[b].Priority = ThreadPriority.BelowNormal;
filteringThreads[b].Name = b.ToString();
filteringThreads[b].Start(filterParams[b]);
}
// and wait for all threads to finish
for (int b = 0; b < noThreads; b++)
{
filteringThreads[b].Join();
filteringThreads[b] = null;
//Console.WriteLine("finished healing thread " + b);
}
foreach (StreamWriter r in filteredReads)
{
if (r != null)
{
r.Close();
}
}
// merge the per-thread histograms
for (int b = 0; b < noThreads; b++)
{
Dictionary <int, int> threadReadDepths = filterParams[b].depthHisto;
foreach (KeyValuePair <int, int> kvp in threadReadDepths)
{
if (readDepths.ContainsKey(kvp.Key))
{
readDepths[kvp.Key] += kvp.Value;
}
else
{
readDepths.Add(kvp.Key, kvp.Value);
}
}
}
} // for a pair of files
StreamWriter histo = new StreamWriter(statsFN);
histo.WriteLine(myProcessNameAndArgs);
histo.WriteLine();
histo.WriteLine("depth\tcount");
int[] depths = readDepths.Keys.ToArray <int>();
int[] counts = readDepths.Values.ToArray <int>();
Array.Sort <int, int>(depths, counts);
for (int i = 0; i < readDepths.Count; i++)
{
histo.WriteLine(depths[i] + "\t" + counts[i]);
}
Console.WriteLine("discarded " + reducedReads + "/" + discardedReads + " of " + totalReads + " reads");
histo.WriteLine("discarded " + reducedReads + "/" + discardedReads + " of " + totalReads + " reads");
histo.Close();
stopMonitor = true;
monitorProgress.Join();
}
static void Main(string[] args)
{
if (args.Length < 2)
{
Console.WriteLine("usage: GenerateMerPairs [-m min] [-t threads] cbtFN readsPattern or file names (" + version + ")");
return;
}
List <string> FNParams = new List <string>(); // the .cbt name and the set of file names or patterns
int noThreads = 1; // no. of healing threads to run in parallel (1 thread is default)
int minLoadReps = 3; // min rep count needed before mer will be loaded into uniqueMers table or saved as a pair
for (int p = 0; p < args.Length; p++)
{
if (args[p][0] == '-')
{
args[p] = args[p].ToLower();
if (args[p] == "-m" || args[p] == "-min")
{
if (!CheckForParamValue(p, args.Length, "minReps number expected after -m|-min"))
{
return;
}
try
{
minLoadReps = Convert.ToInt32(args[p + 1]);
}
catch
{
Console.WriteLine("expected a number for the -m|-min parameter: " + args[p + 1]);
return;
}
p++;
continue;
}
if (args[p] == "-t" || args[p] == "-threads")
{
if (!CheckForParamValue(p, args.Length, "number expected after -t|-threads"))
{
return;
}
try
{
noThreads = Convert.ToInt32(args[p + 1]);
}
catch
{
Console.WriteLine("expected a number for the -t|-threads parameter: " + args[p + 1]);
return;
}
p++;
continue;
}
Console.WriteLine("unrecognised option: " + args[p]);
Console.WriteLine("usage: generateMerPairs [-m min] [-t threads] cbtFN readsPattern or file names (" + version + ")");
return;
}
FNParams.Add(args[p]);
}
if (FNParams.Count < 2)
{
Console.WriteLine("expected a cbt file name and at least one reads file name or pattern");
return;
}
// take the cbt file name from the start of the non-option list
string cbtFN = FNParams[0];
FNParams.RemoveAt(0);
if (FNParams.Count == 0)
{
Console.WriteLine("did not find any reads file names or patterns");
return;
}
string pairsFN = cbtFN.Replace(".cbt", ".prs");
List <string> readsFileNames = new List <string>(FNParams.Count);
List <string> readsFilePaths = new List <string>(FNParams.Count);
foreach (string readsFNP in FNParams)
{
string readsFileName;
string readsFilePath;
GetPathFN(readsFNP, out readsFilePath, out readsFileName);
readsFilePaths.Add(readsFilePath);
readsFileNames.Add(readsFileName);
}
List <string> expandedReadsFNs = new List <string>();
for (int f = 0; f < FNParams.Count; f++)
{
string[] matchedReadsFNs = Directory.GetFiles(readsFilePaths[f], readsFileNames[f], SearchOption.TopDirectoryOnly);
foreach (string matchedReadsFN in matchedReadsFNs)
{
expandedReadsFNs.Add(matchedReadsFN);
}
}
// make sure there aren't any duplicates in the file list (seems to be a bug on the Cherax SGI HPC system and it returns each file name twice)
List <string> distinctReadsFNs = new List <string>();
foreach (string fn in expandedReadsFNs)
{
if (!distinctReadsFNs.Contains(fn))
{
distinctReadsFNs.Add(fn);
}
}
// finally... the set of fully qualified, distinct reads files
string[] readsFNs;
readsFNs = distinctReadsFNs.ToArray();
if (readsFNs.Length == 0)
{
Console.WriteLine("No matching read files found");
return;
}
int noOfReadsFiles = readsFNs.Length;
readsFiles = new StreamReader[noOfReadsFiles];
for (int f = 0; f < noOfReadsFiles; f++)
{
string readsFN = readsFNs[f];
readsFiles[f] = new StreamReader(readsFN);
}
// look at the first file to determine the file format and possible read length
StreamReader testReader = new StreamReader(readsFNs[0]);
char headerChar = (char)testReader.Peek();
if (headerChar == '>')
{
readsFormat = MerStrings.formatFNA;
}
if (headerChar == '@')
{
readsFormat = MerStrings.formatFASTQ;
}
int readLength = 0;
for (int i = 0; i < 20; i++)
{
string nextRead = MerStrings.ReadRead(testReader, readsFormat);
if (nextRead == null)
{
break;
}
int nextLength = nextRead.Length;
if (nextLength > readLength)
{
readLength = nextLength;
}
}
testReader.Close();
// have to able to fit at least two full mers into the read (no overlaps)
if (readLength < 2 * merSize)
{
Console.WriteLine("reads too short to generate pairs: " + readLength);
return;
}
if (!File.Exists(cbtFN))
{
Console.WriteLine(".cbt file not found: " + cbtFN);
return;
}
//string knownPairsFN = "C.sporogenesRaw_25_Copy_1.prs";
//BinaryReader knownPairs = new BinaryReader(File.Open(knownPairsFN, FileMode.Open, FileAccess.Read));
//knownPairs.ReadInt32();
//while (true)
//{
// ulong mer = 0;
// int count = 0;
// try
// {
// mer = knownPairs.ReadUInt64();
// count = knownPairs.ReadInt32();
// goodPairs.Add(mer, count);
// }
// catch
// {
// break;
// }
//}
//knownPairs.Close();
//Console.WriteLine("loaded " + goodPairs.Count + " good mers from " + knownPairsFN);
long loadedUniqueMers = 0;
long loadedTotalMers = 0;
// load the .cbt file into a merTable (either a hash table (small) or a sorted array (large))
MerStrings.LoadCBTFile(cbtFN, minLoadReps, 0, 0, minLoadReps,
out uniqueMers, out merSize, out averageDepth, out loadedUniqueMers, out loadedTotalMers);
if (merSize < merStubSize)
{
Console.WriteLine("mers in .cbt file are shorter than merStub size: " + merSize + " < " + merStubSize);
return;
}
uniquePairs = new MerCollections.MerTables(loadedUniqueMers, noThreads);
// calculate a gap size based on the first read
gap = (readLength - endGuard) / 2 - (merStubSize * 2);
if (gap < minGap)
{
gap = minGap;
}
if (gap > maxGap)
{
gap = maxGap;
}
pairStride = merStubSize + gap + merStubSize;
// start the monitor/synchronising thread
Thread monitorProgress = new Thread(RateReporter);
monitorProgress.Priority = ThreadPriority.AboveNormal;
monitorProgress.Start();
DateTime pairingStart = DateTime.Now;
foreach (string readsFN in readsFNs)
{
Console.WriteLine("Generating pairs from " + readsFN);
StreamReader reads = new StreamReader(readsFN, Encoding.ASCII, false, 1000000);
BufferedReader bufferedReads = new BufferedReader(readsFormat, reads, null);
threadFinished = new EventWaitHandle[noThreads];
int threadNo = 0;
for (int i = 0; i < noThreads; i++)
{
threadFinished[i] = new EventWaitHandle(false, EventResetMode.AutoReset);
}
for (int t = 0; t < noThreads; t++)
{
threadParams workerParam = new threadParams();
workerParam.threadNo = threadNo;
workerParam.bufferedReadsFile = bufferedReads;
ThreadPool.QueueUserWorkItem(new WaitCallback(PairWorker), workerParam);
threadNo++;
}
// and wait for them all to finish
for (int t = 0; t < noThreads; t++)
{
threadFinished[t].WaitOne();
}
}
BinaryWriter pairsFile = new BinaryWriter(File.Open(pairsFN, FileMode.Create, FileAccess.Write));
pairsFile.Write(gap);
for (int pi = 0; pi < uniquePairs.noOfPartitions; pi++)
{
totalPairsGenerated += uniquePairs.repeatedMers[pi].Sort();
}
for (int ti = 0; ti < noThreads; ti++)
{
if (uniquePairs.overflowMers[ti] != null)
{
totalPairsGenerated += uniquePairs.overflowMers[ti].Sort();
}
}
MergeAndWrite(pairsFile, uniquePairs.repeatedMers, uniquePairs.overflowMers);
pairsFile.Close();
StopMonitorThread(monitorProgress);
//Console.WriteLine(totalDeepUnbalancedReads + " deep unbalanced reads");
//Console.WriteLine(totalReadsProcessed + " reads processed");
Console.WriteLine("wrote " + totalPairsWritten + " pairs from " + totalReadsRead + " reads in " + (DateTime.Now - pairingStart).TotalSeconds.ToString("#.0") + "s");
}
