ngs_server 0.1

data/ext/bamtools/src/toolkit/bamtools_stats.cpp

@@ -0,0 +1,286 @@
+// ***************************************************************************
+// bamtools_cpp (c) 2010 Derek Barnett, Erik Garrison
+// Marth Lab, Department of Biology, Boston College
+// ---------------------------------------------------------------------------
+// Last modified: 7 April 2011
+// ---------------------------------------------------------------------------
+// Prints general alignment statistics for BAM file(s).
+// ***************************************************************************
+
+#include "bamtools_stats.h"
+
+#include <api/BamMultiReader.h>
+#include <utils/bamtools_options.h>
+using namespace BamTools;
+
+#include <cmath>
+#include <algorithm>
+#include <functional>
+#include <iostream>
+#include <numeric>
+#include <string>
+#include <vector>
+using namespace std;
+
+// ---------------------------------------------
+// StatsSettings implementation
+
+struct StatsTool::StatsSettings {
+
+    // flags
+    bool HasInput;
+    bool IsShowingInsertSizeSummary;
+
+    // filenames
+    vector<string> InputFiles;
+    
+    // constructor
+    StatsSettings(void)
+        : HasInput(false)
+        , IsShowingInsertSizeSummary(false)
+    { }
+};  
+
+// ---------------------------------------------
+// StatsToolPrivate implementation
+
+struct StatsTool::StatsToolPrivate {
+  
+    // ctor & dtor
+    public:
+        StatsToolPrivate(StatsTool::StatsSettings* _settings);
+        ~StatsToolPrivate(void) { }
+  
+    // 'public' interface
+    public:
+        bool Run(void);
+        
+    // internal methods
+    private:
+        bool CalculateMedian(vector<int>& data, double& median); 
+        void PrintStats(void);
+        void ProcessAlignment(const BamAlignment& al);
+        
+    // data members
+    private:
+        StatsTool::StatsSettings* m_settings;
+        unsigned int m_numReads;
+        unsigned int m_numPaired;
+        unsigned int m_numProperPair;
+        unsigned int m_numMapped;
+        unsigned int m_numBothMatesMapped;
+        unsigned int m_numForwardStrand;
+        unsigned int m_numReverseStrand;
+        unsigned int m_numFirstMate;
+        unsigned int m_numSecondMate;
+        unsigned int m_numSingletons;
+        unsigned int m_numFailedQC;
+        unsigned int m_numDuplicates;
+        vector<int> m_insertSizes;
+};
+
+StatsTool::StatsToolPrivate::StatsToolPrivate(StatsTool::StatsSettings* settings)
+    : m_settings(settings)
+    , m_numReads(0)
+    , m_numPaired(0)
+    , m_numProperPair(0)
+    , m_numMapped(0)
+    , m_numBothMatesMapped(0)
+    , m_numForwardStrand(0)
+    , m_numReverseStrand(0)
+    , m_numFirstMate(0)
+    , m_numSecondMate(0)
+    , m_numSingletons(0)
+    , m_numFailedQC(0)
+    , m_numDuplicates(0)
+{ 
+    m_insertSizes.reserve(100000);
+}
+
+// median is of type double because in the case of even number of data elements,
+// we need to return the average of middle 2 elements
+bool StatsTool::StatsToolPrivate::CalculateMedian(vector<int>& data, double& median) { 
+  
+    // skip if data empty
+    if ( data.empty() ) return false;
+
+    // find middle element
+    size_t middleIndex = data.size() / 2;
+    vector<int>::iterator target = data.begin() + middleIndex;
+    nth_element(data.begin(), target, data.end());
+    
+    // odd number of elements
+    if ( (data.size() % 2) != 0) {
+        median = (double)(*target);
+        return true;
+    }
+    
+    // even number of elements
+    else {
+        double rightTarget = (double)(*target);
+        vector<int>::iterator leftTarget = target - 1;
+        nth_element(data.begin(), leftTarget, data.end());
+        median = (double)((rightTarget+*leftTarget)/2.0);
+        return true;
+    }
+}
+
+// print BAM file alignment stats
+void StatsTool::StatsToolPrivate::PrintStats(void) {
+  
+    cout << endl;
+    cout << "**********************************************" << endl;
+    cout << "Stats for BAM file(s): " << endl;
+    cout << "**********************************************" << endl;
+    cout << endl;
+    cout << "Total reads:       " << m_numReads << endl;
+    cout << "Mapped reads:      " << m_numMapped << "\t(" << ((float)m_numMapped/m_numReads)*100 << "%)" << endl;
+    cout << "Forward strand:    " << m_numForwardStrand << "\t(" << ((float)m_numForwardStrand/m_numReads)*100 << "%)" << endl;
+    cout << "Reverse strand:    " << m_numReverseStrand << "\t(" << ((float)m_numReverseStrand/m_numReads)*100 << "%)" << endl;
+    cout << "Failed QC:         " << m_numFailedQC << "\t(" << ((float)m_numFailedQC/m_numReads)*100 << "%)" << endl;
+    cout << "Duplicates:        " << m_numDuplicates << "\t(" << ((float)m_numDuplicates/m_numReads)*100 << "%)" << endl;
+    cout << "Paired-end reads:  " << m_numPaired << "\t(" << ((float)m_numPaired/m_numReads)*100 << "%)" << endl;
+    
+    if ( m_numPaired != 0 ) {
+        cout << "'Proper-pairs':    " << m_numProperPair << "\t(" << ((float)m_numProperPair/m_numPaired)*100 << "%)" << endl;
+        cout << "Both pairs mapped: " << m_numBothMatesMapped << "\t(" << ((float)m_numBothMatesMapped/m_numPaired)*100 << "%)" << endl;
+        cout << "Read 1:            " << m_numFirstMate << endl;
+        cout << "Read 2:            " << m_numSecondMate << endl;
+        cout << "Singletons:        " << m_numSingletons << "\t(" << ((float)m_numSingletons/m_numPaired)*100 << "%)" << endl;
+    }
+    
+    if ( m_settings->IsShowingInsertSizeSummary ) {
+      
+        double avgInsertSize = 0.0;
+        if ( !m_insertSizes.empty() ) {
+            avgInsertSize = ( accumulate(m_insertSizes.begin(), m_insertSizes.end(), 0.0) / (double)m_insertSizes.size() );
+            cout << "Average insert size (absolute value): " << avgInsertSize << endl;
+        }
+        
+        double medianInsertSize = 0.0;
+        if ( CalculateMedian(m_insertSizes, medianInsertSize) )
+            cout << "Median insert size (absolute value): " << medianInsertSize << endl;
+    }
+    cout << endl;
+}
+
+// use current input alignment to update BAM file alignment stats
+void StatsTool::StatsToolPrivate::ProcessAlignment(const BamAlignment& al) {
+  
+    // increment total alignment counter
+    ++m_numReads;
+    
+    // incrememt counters for pairing-independent flags
+    if ( al.IsDuplicate() ) ++m_numDuplicates;
+    if ( al.IsFailedQC()  ) ++m_numFailedQC;
+    if ( al.IsMapped()    ) ++m_numMapped;
+    
+    // increment strand counters
+    if ( al.IsReverseStrand() ) 
+        ++m_numReverseStrand;
+    else 
+        ++m_numForwardStrand;
+    
+    // if alignment is paired-end
+    if ( al.IsPaired() ) {
+      
+        // increment PE counter
+        ++m_numPaired;
+      
+        // increment first mate/second mate counters
+        if ( al.IsFirstMate()  ) ++m_numFirstMate;
+        if ( al.IsSecondMate() ) ++m_numSecondMate;
+        
+        // if alignment is mapped, check mate status
+        if ( al.IsMapped() ) {
+            // if mate mapped
+            if ( al.IsMateMapped() ) 
+                ++m_numBothMatesMapped;
+            // else singleton
+            else 
+                ++m_numSingletons;
+        }
+        
+        // check for explicit proper pair flag
+        if ( al.IsProperPair() ) ++m_numProperPair;
+        
+        // store insert size for first mate 
+        if ( m_settings->IsShowingInsertSizeSummary && al.IsFirstMate() && (al.InsertSize != 0) ) {
+            int insertSize = abs(al.InsertSize);
+            m_insertSizes.push_back( insertSize );
+        }
+    }
+}
+
+bool StatsTool::StatsToolPrivate::Run() {
+  
+    // set to default input if none provided
+    if ( !m_settings->HasInput )
+        m_settings->InputFiles.push_back(Options::StandardIn());
+
+    // open the BAM files
+    BamMultiReader reader;
+    if ( !reader.Open(m_settings->InputFiles) ) {
+        cerr << "bamtools stats ERROR: could not open input BAM file(s)... Aborting." << endl;
+        reader.Close();
+        return false;
+    }
+    
+    // plow through alignments, keeping track of stats
+    BamAlignment al;
+    while ( reader.GetNextAlignmentCore(al) )
+        ProcessAlignment(al);
+    reader.Close();
+    
+    // print stats & exit
+    PrintStats();
+    return true; 
+}
+
+// ---------------------------------------------
+// StatsTool implementation
+
+StatsTool::StatsTool(void)
+    : AbstractTool()
+    , m_settings(new StatsSettings)
+    , m_impl(0)
+{
+    // set program details
+    Options::SetProgramInfo("bamtools stats", "prints general alignment statistics", "[-in <filename> -in <filename> ...] [statsOptions]");
+    
+    // set up options 
+    OptionGroup* IO_Opts = Options::CreateOptionGroup("Input & Output");
+    Options::AddValueOption("-in", "BAM filename", "the input BAM file", "", m_settings->HasInput,  m_settings->InputFiles,  IO_Opts, Options::StandardIn());
+    
+    OptionGroup* AdditionalOpts = Options::CreateOptionGroup("Additional Stats");
+    Options::AddOption("-insert", "summarize insert size data", m_settings->IsShowingInsertSizeSummary, AdditionalOpts);
+}
+
+StatsTool::~StatsTool(void) {
+
+    delete m_settings;
+    m_settings = 0;
+    
+    delete m_impl;
+    m_impl = 0;
+}
+
+int StatsTool::Help(void) {
+    Options::DisplayHelp();
+    return 0;
+}
+
+int StatsTool::Run(int argc, char* argv[]) {
+  
+    // parse command line arguments
+    Options::Parse(argc, argv, 1);
+    
+    // initialize StatsTool with settings
+    m_impl = new StatsToolPrivate(m_settings);
+    
+    // run StatsTool, return success/fail
+    if ( m_impl->Run() )
+        return 0;
+    else
+        return 1;
+}

data/ext/bamtools/src/toolkit/bamtools_stats.h

@@ -0,0 +1,37 @@
+// ***************************************************************************
+// bamtools_stats.h (c) 2010 Derek Barnett, Erik Garrison
+// Marth Lab, Department of Biology, Boston College
+// ---------------------------------------------------------------------------
+// Last modified: 7 April 2011
+// ---------------------------------------------------------------------------
+// Prints general statistics for a single BAM file
+// ***************************************************************************
+
+#ifndef BAMTOOLS_STATS_H
+#define BAMTOOLS_STATS_H
+
+#include "bamtools_tool.h"
+
+namespace BamTools {
+  
+class StatsTool : public AbstractTool {
+  
+    public:
+        StatsTool(void);
+        ~StatsTool(void);
+  
+    public:
+        int Help(void);
+        int Run(int argc, char* argv[]); 
+        
+    private:
+        struct StatsSettings;
+        StatsSettings* m_settings;
+        
+        struct StatsToolPrivate;
+        StatsToolPrivate* m_impl;
+};
+  
+} // namespace BamTools
+
+#endif // BAMTOOLS_STATS_H

data/ext/bamtools/src/toolkit/bamtools_tool.h

@@ -0,0 +1,35 @@
+// ***************************************************************************
+// bamtools_tool.h (c) 2010 Derek Barnett, Erik Garrison
+// Marth Lab, Department of Biology, Boston College
+// ---------------------------------------------------------------------------
+// Last modified: 2 June 2010
+// ---------------------------------------------------------------------------
+// Base class for all other BamTools sub-tools
+// All derived classes must provide Help() and Run() methods
+// ***************************************************************************
+
+#ifndef BAMTOOLS_ABSTRACTTOOL_H
+#define BAMTOOLS_ABSTRACTTOOL_H
+
+#include <string>
+
+namespace BamTools { 
+  
+class AbstractTool {
+  
+    public:
+        AbstractTool(void) { }
+        virtual ~AbstractTool(void) { }
+
+    public:
+        virtual int Help(void) =0;
+        virtual int Run(int argc, char* argv[]) =0; 
+
+    // derived classes should also provide:
+    // static std::string Description(void);
+    // static std::String Name(void);
+};
+  
+} // namespace BamTools
+
+#endif // BAMTOOLS_ABSTRACTTOOL_H

data/ext/bamtools/src/toolkit/bamtools_version.h

@@ -0,0 +1,20 @@
+// ***************************************************************************
+// bamtools_version.h.in (c) 2010 Derek Barnett
+// Marth Lab, Department of Biology, Boston College
+// ---------------------------------------------------------------------------
+// Last modified: 19 November 2010 (DB)
+// ---------------------------------------------------------------------------
+// Provides version information for the BamTools toolkit.
+// ***************************************************************************
+
+#ifndef BAMTOOLS_VERSION_H
+#define BAMTOOLS_VERSION_H
+
+// CMake uses this file as a template to generate "bamtools_version.h".
+// These constants are defined to match the variables set in the build system. 
+#define BAMTOOLS_VERSION_MAJOR 1
+#define BAMTOOLS_VERSION_MINOR 0
+#define BAMTOOLS_VERSION_BUILD 2
+
+#endif // BAMTOOLS_VERSION_H
+

data/ext/bamtools/src/toolkit/bamtools_version.h.in

@@ -0,0 +1,20 @@
+// ***************************************************************************
+// bamtools_version.h.in (c) 2010 Derek Barnett
+// Marth Lab, Department of Biology, Boston College
+// ---------------------------------------------------------------------------
+// Last modified: 19 November 2010 (DB)
+// ---------------------------------------------------------------------------
+// Provides version information for the BamTools toolkit.
+// ***************************************************************************
+
+#ifndef BAMTOOLS_VERSION_H
+#define BAMTOOLS_VERSION_H
+
+// CMake uses this file as a template to generate "bamtools_version.h".
+// These constants are defined to match the variables set in the build system. 
+#define BAMTOOLS_VERSION_MAJOR @BamTools_VERSION_MAJOR@
+#define BAMTOOLS_VERSION_MINOR @BamTools_VERSION_MINOR@
+#define BAMTOOLS_VERSION_BUILD @BamTools_VERSION_BUILD@
+
+#endif // BAMTOOLS_VERSION_H
+

data/ext/bamtools/src/toolkit/cmake_install.cmake

@@ -0,0 +1,52 @@
+# Install script for directory: /Users/chase/Desktop/tmp_workspace/ngs_server/ext/bamtools/src/toolkit
+
+# Set the install prefix
+IF(NOT DEFINED CMAKE_INSTALL_PREFIX)
+  SET(CMAKE_INSTALL_PREFIX "/usr/local")
+ENDIF(NOT DEFINED CMAKE_INSTALL_PREFIX)
+STRING(REGEX REPLACE "/$" "" CMAKE_INSTALL_PREFIX "${CMAKE_INSTALL_PREFIX}")
+
+# Set the install configuration name.
+IF(NOT DEFINED CMAKE_INSTALL_CONFIG_NAME)
+  IF(BUILD_TYPE)
+    STRING(REGEX REPLACE "^[^A-Za-z0-9_]+" ""
+           CMAKE_INSTALL_CONFIG_NAME "${BUILD_TYPE}")
+  ELSE(BUILD_TYPE)
+    SET(CMAKE_INSTALL_CONFIG_NAME "Release")
+  ENDIF(BUILD_TYPE)
+  MESSAGE(STATUS "Install configuration: \"${CMAKE_INSTALL_CONFIG_NAME}\"")
+ENDIF(NOT DEFINED CMAKE_INSTALL_CONFIG_NAME)
+
+# Set the component getting installed.
+IF(NOT CMAKE_INSTALL_COMPONENT)
+  IF(COMPONENT)
+    MESSAGE(STATUS "Install component: \"${COMPONENT}\"")
+    SET(CMAKE_INSTALL_COMPONENT "${COMPONENT}")
+  ELSE(COMPONENT)
+    SET(CMAKE_INSTALL_COMPONENT)
+  ENDIF(COMPONENT)
+ENDIF(NOT CMAKE_INSTALL_COMPONENT)
+
+IF(NOT CMAKE_INSTALL_COMPONENT OR "${CMAKE_INSTALL_COMPONENT}" STREQUAL "Unspecified")
+  FILE(INSTALL DESTINATION "${CMAKE_INSTALL_PREFIX}/bin" TYPE EXECUTABLE FILES
+    "/Users/chase/Desktop/tmp_workspace/ngs_server/ext/bamtools/bin/bamtools-1.0.2"
+    "/Users/chase/Desktop/tmp_workspace/ngs_server/ext/bamtools/bin/bamtools"
+    )
+  FOREACH(file
+      "$ENV{DESTDIR}${CMAKE_INSTALL_PREFIX}/bin/bamtools-1.0.2"
+      "$ENV{DESTDIR}${CMAKE_INSTALL_PREFIX}/bin/bamtools"
+      )
+    IF(EXISTS "${file}" AND
+       NOT IS_SYMLINK "${file}")
+      EXECUTE_PROCESS(COMMAND "/usr/bin/install_name_tool"
+        -change "/Users/chase/Desktop/tmp_workspace/ngs_server/ext/bamtools/lib/libbamtools-utils.1.0.2.dylib" "libbamtools-utils.1.0.2.dylib"
+        -change "/Users/chase/Desktop/tmp_workspace/ngs_server/ext/bamtools/lib/libbamtools.1.0.2.dylib" "libbamtools.1.0.2.dylib"
+        -change "/Users/chase/Desktop/tmp_workspace/ngs_server/ext/bamtools/lib/libjsoncpp.1.0.0.dylib" "libjsoncpp.1.0.0.dylib"
+        "${file}")
+      IF(CMAKE_INSTALL_DO_STRIP)
+        EXECUTE_PROCESS(COMMAND "/usr/bin/strip" "${file}")
+      ENDIF(CMAKE_INSTALL_DO_STRIP)
+    ENDIF()
+  ENDFOREACH()
+ENDIF(NOT CMAKE_INSTALL_COMPONENT OR "${CMAKE_INSTALL_COMPONENT}" STREQUAL "Unspecified")
+

data/ext/bamtools/src/utils/CMakeFiles/BamTools-utils.dir/CXX.includecache

@@ -0,0 +1,250 @@
+#IncludeRegexLine: ^[ 	]*#[ 	]*(include|import)[ 	]*[<"]([^">]+)([">])
+
+#IncludeRegexScan: ^.*$
+
+#IncludeRegexComplain: ^$
+
+#IncludeRegexTransform: 
+
+/Users/chase/Desktop/tmp_workspace/ngs_server/ext/bamtools/src/utils/bamtools_fasta.cpp
+utils/bamtools_fasta.h
+-
+cstdio
+-
+cstdlib
+-
+cstring
+-
+fstream
+-
+iostream
+-
+sstream
+-
+vector
+-
+
+/Users/chase/Desktop/tmp_workspace/ngs_server/ext/bamtools/src/utils/bamtools_options.cpp
+utils/bamtools_options.h
+-
+cstdio
+-
+cstdlib
+-
+cstring
+-
+iomanip
+-
+sstream
+-
+
+/Users/chase/Desktop/tmp_workspace/ngs_server/ext/bamtools/src/utils/bamtools_pileup_engine.cpp
+utils/bamtools_pileup_engine.h
+-
+iostream
+-
+
+/Users/chase/Desktop/tmp_workspace/ngs_server/ext/bamtools/src/utils/bamtools_utilities.cpp
+api/BamMultiReader.h
+-
+api/BamReader.h
+-
+utils/bamtools_utilities.h
+-
+algorithm
+-
+cstdlib
+-
+cstring
+-
+fstream
+-
+iostream
+-
+sstream
+-
+
+src/api/BamAlignment.h
+api/api_global.h
+-
+api/BamAux.h
+-
+string
+-
+vector
+-
+
+src/api/BamAux.h
+api/api_global.h
+-
+fstream
+-
+iostream
+-
+string
+-
+vector
+-
+
+src/api/BamIndex.h
+api/api_global.h
+-
+api/BamAux.h
+-
+string
+-
+
+src/api/BamMultiReader.h
+api/api_global.h
+-
+api/BamReader.h
+-
+map
+-
+sstream
+-
+string
+-
+utility
+-
+
+src/api/BamReader.h
+api/api_global.h
+-
+api/BamAlignment.h
+-
+api/BamIndex.h
+-
+api/SamHeader.h
+-
+string
+-
+
+src/api/SamHeader.h
+api/api_global.h
+-
+api/SamProgramChain.h
+-
+api/SamReadGroupDictionary.h
+-
+api/SamSequenceDictionary.h
+-
+string
+-
+vector
+-
+
+src/api/SamProgram.h
+api/api_global.h
+src/api/api/api_global.h
+string
+-
+
+src/api/SamProgramChain.h
+api/api_global.h
+-
+api/SamProgram.h
+-
+string
+-
+vector
+-
+
+src/api/SamReadGroup.h
+api/api_global.h
+src/api/api/api_global.h
+string
+-
+
+src/api/SamReadGroupDictionary.h
+api/api_global.h
+-
+api/SamReadGroup.h
+-
+string
+-
+vector
+-
+
+src/api/SamSequence.h
+api/api_global.h
+-
+string
+-
+
+src/api/SamSequenceDictionary.h
+api/api_global.h
+-
+api/SamSequence.h
+-
+string
+-
+map
+-
+vector
+-
+
+src/api/api_global.h
+shared/bamtools_global.h
+src/api/shared/bamtools_global.h
+
+src/shared/bamtools_global.h
+stdint.h
+-
+
+src/utils/bamtools_fasta.h
+utils/utils_global.h
+-
+string
+-
+
+src/utils/bamtools_options.h
+utils/bamtools_variant.h
+-
+utils/utils_global.h
+-
+map
+-
+string
+-
+vector
+-
+stdint.h
+-
+
+src/utils/bamtools_pileup_engine.h
+api/BamAlignment.h
+-
+utils/utils_global.h
+-
+vector
+-
+
+src/utils/bamtools_utilities.h
+api/BamAux.h
+-
+utils/utils_global.h
+-
+cassert
+-
+stdexcept
+-
+string
+-
+vector
+-
+
+src/utils/bamtools_variant.h
+utils/utils_global.h
+-
+stdexcept
+-
+string
+-
+typeinfo
+-
+
+src/utils/utils_global.h
+shared/bamtools_global.h
+src/utils/shared/bamtools_global.h
+