Major refactoring to have a map-based hg38 index/converter.

* Added Makefile to replace build scripts.

Signed-off-by: Philip R. Kensche <p.kensche@dkfz.de>

Makefile

@@ -0,0 +1,104 @@
+# Compiler
+CXX = x86_64-conda_cos6-linux-gnu-g++
+
+INCLUDE_DIR = ./include
+BUILD_DIR = ./build
+SRC_DIR = ./src
+
+
+INCLUDE_FLAGS= -Iinclude -I$$CONDA_PREFIX/include
+LIBDIR_FLAGS = -L$$CONDA_PREFIX/lib
+
+# Compiler flags
+CXXFLAGS = -O3 -Wall -Wextra -static -static-libgcc -static-libstdc++ -flto  -c -fmessage-length=0 -Wno-attributes
+ifeq ($(STATIC), "true")
+	CXXFLAGS := -static -static-libgcc -static-libstdc++ $(CXXFLAGS)
+endif
+CXXFLAGS := $(LIBDIR_FLAGS) -std=c++17 $(INCLUDE_FLAGS) $(CXXFLAGS)
+
+# Source files
+SOURCES = $(wildcard src/*.cpp)
+
+# Object files should have .o instead of .cpp.
+OBJECTS = $(SOURCES:.cpp=.o)
+
+# Binaries
+BINARIES = sophia sophiaAnnotate sophiaMref
+
+# Default rule
+all: $(BINARIES)
+
+$(BUILD_DIR):
+	mkdir -p $@
+
+# Rule for object files
+$(BUILD_DIR)/%.o: %.cpp | $(BUILD_DIR)
+	$(CXX) $(INCLUDE_FLAGS) $(LIBDIR_FLAGS) $(CXXFLAGS) -c $< -o $@
+
+download_strtk: include/strtk.hpp
+	wget -c https://github.com/ArashPartow/strtk/raw/master/strtk.hpp -O include/strtk.hpp
+
+vpath %.h includes
+vpath %.cpp src
+
+# Rule for sophia
+sophia: $(BUILD_DIR)/Alignment.o \
+        $(BUILD_DIR)/Breakpoint.o \
+        $(BUILD_DIR)/ChosenBp.o \
+        $(BUILD_DIR)/ChrConverter.o \
+        $(BUILD_DIR)/Hg37ChrConverter.o \
+        $(BUILD_DIR)/Hg38ChrConverter.o \
+        $(BUILD_DIR)/SamSegmentMapper.o \
+        $(BUILD_DIR)/Sdust.o \
+        $(BUILD_DIR)/SuppAlignment.o \
+        $(BUILD_DIR)/HelperFunctions.o \
+        $(BUILD_DIR)/GlobalAppConfig.o \
+        $(BUILD_DIR)/sophia.o \
+        download_strtk
+	$(CXX) $(LIBDIR_FLAGS) -lboost_program_options -o $@ $^
+
+# Rule for sophiaAnnotate
+sophiaAnnotate: $(BUILD_DIR)/AnnotationProcessor.o \
+				$(BUILD_DIR)/Breakpoint.o \
+				$(BUILD_DIR)/BreakpointReduced.o \
+				$(BUILD_DIR)/ChrConverter.o \
+				$(BUILD_DIR)/Hg37ChrConverter.o \
+				$(BUILD_DIR)/Hg38ChrConverter.o \
+				$(BUILD_DIR)/DeFuzzier.o \
+				$(BUILD_DIR)/GermlineMatch.o \
+				$(BUILD_DIR)/MrefEntry.o \
+				$(BUILD_DIR)/MrefEntryAnno.o \
+				$(BUILD_DIR)/MrefMatch.o \
+				$(BUILD_DIR)/SuppAlignment.o \
+				$(BUILD_DIR)/SuppAlignmentAnno.o \
+				$(BUILD_DIR)/SvEvent.o \
+				$(BUILD_DIR)/HelperFunctions.o \
+				$(BUILD_DIR)/GlobalAppConfig.o \
+				$(BUILD_DIR)/sophiaAnnotate.o \
+				download_strtk
+	$(CXX) $(LIBDIR_FLAGS) -lz -lboost_system -lboost_iostreams $(CXXFLAGS) -o $@ $^
+
+# Rule for sophiaMref
+sophiaMref: $(BUILD_DIR)/GlobalAppConfig.o \
+			$(BUILD_DIR)/ChrConverter.o \
+			$(BUILD_DIR)/Hg37ChrConverter.o \
+			$(BUILD_DIR)/Hg38ChrConverter.o \
+			$(BUILD_DIR)/HelperFunctions.o \
+			$(BUILD_DIR)/SuppAlignment.o \
+			$(BUILD_DIR)/SuppAlignmentAnno.o \
+			$(BUILD_DIR)/MrefEntry.o \
+			$(BUILD_DIR)/MrefEntryAnno.o \
+			$(BUILD_DIR)/MrefMatch.o \
+			$(BUILD_DIR)/MasterRefProcessor.o \
+			$(BUILD_DIR)/Breakpoint.o \
+			$(BUILD_DIR)/BreakpointReduced.o \
+			$(BUILD_DIR)/GermlineMatch.o \
+			$(BUILD_DIR)/DeFuzzier.o \
+			$(BUILD_DIR)/sophiaMref.o \
+			download_strtk
+	$(CXX) $(LIBDIR_FLAGS) -lz -lboost_system -lboost_iostreams -lboost_program_options $(CXXFLAGS) -o $@ $^
+
+# Rule for clean
+.PHONY: clean
+clean:
+	rm -f $(OBJECTS) $(BINARIES)

README.md

@@ -25,40 +25,47 @@ You can cite Sophia as follows:
     Umut Toprak (2019).
     DOI 10.11588/heidok.000274296
 
+### Tools
+
+* `sophia` - The main tool for SV calling. It takes a BAM file as input and outputs a list of SVs in mref format.
+* `sophiaAnnotate` - Tool for annotating SVs with gene information. It reads in an mref file created by `sophiaMref` and annotates the SVs in the input file with gene information.
+* `sophiaMref` - The `sophiaMref` tool processes a list of gzipped control bed files and generates a reference that can be used by `sophiaAnnotate` for annotating structural variants with gene information.
+
+For instructions on commandline parameters, invoke the tool with `--help`.
+
 ## Runtime Dependencies
 
-The only dependency is Boost 1.70.0 (currently). E.g. you can do
+The only dependency is Boost 1.82.0 (currently). E.g. you can do
 
 ```bash
-conda create -n sophia boost=1.70.0
+conda create -n sophia boost=1.82.0
 ```
 
 ## Building
 
-### Build-time Dependencies
-
-* g++ >= 7
-* Boost 1.70.0
-
 ### Dynamic Build
 
 With Conda you can do
 
 ```bash
-conda create -n sophia gxx_linux-64=8 boost=1.70.0
+conda create -n sophia gxx_linux-64=8 boost=1.82.0
 ```
 
-to create an environment to build the `sophia` and `sophiaAnnotate` binaries.
+to create an environment to build the SOPHIA binaries binaries.
 
 To build you need to do
 
 ```bash
 source activate sophia
-cd Release_sophia
-build-sophia.sh
+
+cd Release_sophiaMref
+./build-sophiaMref.sh
+
+cd ../Release_sophia
+./build-sophia.sh
 
 cd ../Release_sophiaAnnotate
-build-sophiaAnnotate.sh
+./build-sophiaAnnotate.sh
 ```
 
 Note that the build-scripts are for when you manage your dependencies with Conda.

include/ChrConverter.h

@@ -26,51 +26,52 @@
 
 namespace sophia {
 
-    using namespace std;
+    // These two are only to make the code clearer, but are not type checked. There are no opaque
+    // or strongly type-checked typedefs in C++17.
+    typedef size_t ChrIndex;
+    typedef size_t CompressedMrefIndex;
 
-    /** ChrConverter contains information the names of chromosomes in an assembly. */
-    class ChrConverter {
-      protected:
-
-        /** The constructor should be used to initialize the fields from subclasses. It does
-            additional checks of the dimensions of the input vectors. */
-        ChrConverter(const vector<string>& indexToChr,
-                     const vector<string>& indexToChrCompressedMref,
-                     const vector<int>& chrSizesCompressedMref,
-                     const vector<int>& indexConverter);
+    /** ChrConverter manages information on chromosomes names, sizes, and index positions in
+        data arrays.
 
+        This probably needs a redesign. The current situation is just an intermediate step away
+        from the former completely procedural implementation that heavily leaked implementation
+        details into the calling code and was highly tuned but very inflexible. */
+    class ChrConverter {
 
       public:
 
         virtual ~ChrConverter();
 
         /** The name of the assembly. */
-        static const string assembly_name;
+        static const std::string assemlyName;
 
-        /** Mapping indices to chromosome names. */
-        const vector<string> indexToChr;
+        /** Number of chromosomes. */
+        virtual int nChromosomes() const = 0;
 
-        /** Mapping indices to chromosome names for compressed mref files. */
-        const vector<string> indexToChrCompressedMref;
+        /** Number of compressed mref chromosomes. */
+        virtual int nChromosomesCompressedMref() const = 0;
 
-        /** Chromosome sizes in base pairs. */
-        const vector<int> chrSizesCompressedMref;
+        /** Map an index position to a chromosome name. */
+        virtual std::string indexToChrName(ChrIndex index) const = 0;
 
-        /** Mapping chromosome names to indices. */
-        const vector<int> indexConverter;
+        /** Map an index position to a chromosome name for compressed mref files. */
+        virtual std::string indexToChrNameCompressedMref(CompressedMrefIndex index) const = 0;
 
-        /** Parse chromosome index.  It takes a position in a character stream, and translates the
-            following character(s) into index positions (using ChrConverter::indexToChr). */
-        virtual int readChromosomeIndex(string::const_iterator startIt, char stopChar) const = 0;
+        /** Map the compressed mref index to the uncompressed mref index. */
+        virtual ChrIndex compressedMrefIndexToIndex(CompressedMrefIndex index) const = 0;
 
-        size_t n_chromosomes() {
-            return indexToChr.size();
-        };
+        /** Map compressed mref index to chromosome size. */
+        virtual int chrSizeCompressedMref(CompressedMrefIndex index) const = 0;
 
-        size_t n_chromosomes_compressed_mref() {
-            return indexToChrCompressedMref.size();
-        };
+        /** Map a chromosome name to an index position. */
+        virtual ChrIndex chrNameToIndex(std::string chrName) const = 0;
 
+        /** Parse chromosome index. It takes a position in a character stream, and translates the
+            following character(s) into index positions (using ChrConverter::indexToChr).
+            If the name cannot be parsed, throws a domain_error exception. */
+        virtual ChrIndex parseChrAndReturnIndex(std::string::const_iterator startIt,
+                                                char stopChar) const = 0;
 
     };
 

include/GlobalAppConfig.h

@@ -38,12 +38,12 @@ namespace sophia {
 
       protected:
 
-        GlobalAppConfig(unique_ptr<ChrConverter const> chrConverter);
+        GlobalAppConfig(std::unique_ptr<ChrConverter const> chrConverter);
 
         ~GlobalAppConfig();
 
         /** The chromosome converter. */
-        const unique_ptr<ChrConverter const> chrConverter;
+        const std::unique_ptr<ChrConverter const> chrConverter;
 
       public:
 
@@ -56,7 +56,7 @@ namespace sophia {
         void operator=(const GlobalAppConfig &) = delete;
 
         /** Factory method. */
-        static GlobalAppConfig &init(unique_ptr<ChrConverter const> chrConverter);
+        static GlobalAppConfig &init(std::unique_ptr<ChrConverter const> chrConverter);
 
         /** Getter. */
         static const GlobalAppConfig &getInstance();

include/Hg37ChrConverter.h

@@ -26,16 +26,59 @@
 
 namespace sophia {
 
-    using namespace std;
-
+    /** Hard-coded chromosome converter for hg37. This tries to encapsulate the implementation
+        details of the original version. */
     class Hg37ChrConverter: public ChrConverter {
+      protected:
+
+        /** The constructor does additional checks of the dimensions of the input vectors. */
+        Hg37ChrConverter(const std::vector<std::string>& indexToChr,
+                         const std::vector<std::string>& indexToChrCompressedMref,
+                         const std::vector<CompressedMrefIndex>& chrSizesCompressedMref,
+                         const std::vector<ChrIndex>& indexConverter);
+
+        /** Mapping indices to chromosome names. */
+        const std::vector<std::string> indexToChr;
+
+        /** Mapping indices to chromosome names for compressed mref indices. */
+        const std::vector<std::string> indexToChrCompressedMref;
+
+        /** Chromosome sizes in base pairs, only for compressed mref chromosomes. */
+        const std::vector<CompressedMrefIndex> chrSizesCompressedMref;
+
+        /** Mapping compressed mref indices names to indices. */
+        const std::vector<ChrIndex> indexConverter;
+
       public:
 
-        static const string assembly_name;
+        static const std::string assemblyName;
 
         Hg37ChrConverter();
 
-        int readChromosomeIndex(string::const_iterator startIt, char stopChar) const;
+        /** Return the number of chromosomes. */
+        int nChromosomes() const;
+
+        /** Number of compressed mref chromosomes. */
+        int nChromosomesCompressedMref() const;
+
+        /** Map an index position to a chromosome name. */
+        std::string indexToChrName(ChrIndex index) const;
+
+        /** Map an index position to a chromosome name for compressed mref files. */
+        std::string indexToChrNameCompressedMref(CompressedMrefIndex index) const;
+
+        /** Map the compressed mref index to the uncompressed mref index. */
+        ChrIndex compressedMrefIndexToIndex(CompressedMrefIndex index) const;
+
+        /** Map compressed mref index to chromosome size. */
+        int chrSizeCompressedMref(CompressedMrefIndex index) const;
+
+        /** Map a chromosome name to an index position for compressed mref files. */
+        CompressedMrefIndex chrNameToIndexCompressedMref(std::string chrName) const;
+
+        /** Parse chromosome name given a iterator (start) and termination character.
+            Validate against the pre-declared chromosome names. */
+        ChrIndex parseChrAndReturnIndex(std::string::const_iterator startIt, char stopChar) const;
 
     };