initial commit

Makefile

@@ -0,0 +1,19 @@
+all: bin/Red.o bin/meshclust2
+
+bin/Red.o:
+	mkdir -p bin
+	mkdir -p bin/exception
+	mkdir -p bin/nonltr
+	mkdir -p bin/utility
+	$(MAKE) -C src
+bin/meshclust2: bin/Red.o
+	$(MAKE) -C src/cluster
+	cp src/cluster/meshclust2 bin
+
+clean:
+	$(MAKE) clean -C src
+	$(MAKE) clean -C src/cluster
+	$(RM) -r bin
+
+rebuild: clean all
+.PHONY: all clean

README

@@ -0,0 +1,74 @@
+MeShClust2
+Release version
+
+Requirements: g++ 4.9.1 or later, requires Homebrew on Mac OS X
+
+Compilation using g++ (homebrew) and GNU Make on Mac OS X
+CXX=g++-7 make
+
+see: https://stackoverflow.com/questions/29057437/compile-openmp-programs-with-gcc-compiler-on-os-x-yosemite
+
+
+Linux/Unix compilation:
+make
+
+Usage: bin/meshclust2 --id 0.x [OPTIONS] *.fasta
+
+--id  The most important parameter, --id, controls the identity cutoff of the sequences.
+      Needs to be between 0 and 1.
+      If it is not specified, an identity of 0.9 is used.
+
+--kmer decides the size of the kmers. It is by default automatically decided by average sequence length,
+       but if provided, MeShClust can speed up a little by not having to find the largest sequence length.
+       Increasing kmer size can increase accuracy, but increases memory consumption.
+
+--mut-type {single, both, nonsingle-typical, nonsingle-all, all-but-reversion, all-but-translocation}
+	   changes the mutation generation algorithm. By default, "single" is used, utilizing only
+	   single point mutations. On low identity data sets, "both", which includes single mutations
+	   and block mutations, is preferable. The option "nonsingle-typical" uses only block mutations,
+	   disallowing single point mutations. Other options include "all", which includes single,
+	   block, and nontypical mutations translocation and reversion.
+
+--feat    determines the combinations of features to be used. By default, "fast" allows 9 fast combinations
+	  to be selected from. "slow" adds 2 slower features which include logarithm based features,
+	  and "extraslow" includes 33 total features used in a previous study.
+
+--min-feat (default 3) sets the minimum feature pairs to be used. If set to 2, at least 2 feature pairs
+	   will be used. Recall that features include pairwise combinations of the "feat" option.
+
+--max-feat (default 5) sets the maximum feature pairs to be used. Diminishing returns appears quickly,
+	   so a very large maximum is not advised.
+
+--sample selects the total number of sequences used for both training and testing.
+	 300 is the default value. Each sequence generates 10 synthetic mutants.
+	 That is, --sample 300 provides 3000 training pairs and 3000 testing pairs.
+
+--min-id  (default 0.35) sets the lower bound for mutation identity scores to be calculated. Shouldn't need
+	  to be set normally, as lower identites take much longer, especially with single mutations only.
+
+--threads sets the number of threads to be used. By default OpenMP uses the number of available cores
+	  on your machine, but this parameter overwrites that.
+
+--output specifies the output file, in CD-HIT's CLSTR format, described below:
+	 A '>Cluster ' followed by an increasing index designates a cluster.
+	 Otherwise, the sequence is printed out.
+	 A '*' at the end of a sequence designates the center of the cluster.
+	 An example of a small data set:
+
+	 >Cluster 0
+	 0       993nt, >seq128 template_6... *
+	 >Cluster 1
+	 0       1043nt, >seq235 template_10...
+	 1       1000nt, >seq216 template_10... *
+	 2       1015nt, >seq237 template_10...
+
+
+--delta decides how many clusters are looked around in the final clustering stage.
+	Increasing it creates more accuracy, but takes more time. Default value is 5.
+
+--iterations specifies how many iterations in the final stage of merging are done until convergence.
+	     Default value is 15.
+
+
+
+If the argument is not listed here, it is interpreted as an input (FASTA format) file.

src/Makefile

@@ -0,0 +1,175 @@
+# CXX = /usr/bin/c++
+CXX ?= g++
+
+CXXFLAGS =	-O3 -g -fmessage-length=0 -Wall -march=native -std=c++11
+
+#
+# Objects
+#
+
+ORed  = ../bin/Red.o
+
+# Exception
+OInvalidInputException  = ../bin/exception/InvalidInputException.o
+OInvalidStateException  = ../bin/exception/InvalidStateException.o
+OFileDoesNotExistException  = ../bin/exception/FileDoesNotExistException.o
+OInvalidOrderOfOperationsException  = ../bin/exception/InvalidOrderOfOperationsException.o
+OInvalidScoreException  = ../bin/exception/InvalidScoreException.o
+OInvalidOperationException = ../bin/exception/InvalidOperationException.o
+
+# Utility
+OUtil = ../bin/utility/Util.o
+OLocation = ../bin/utility/Location.o
+OEmptyLocation = ../bin/utility/EmptyLocation.o
+OLCSLen = ../bin/utility/LCSLen.o
+OAffineId = ../bin/utility/AffineId.o
+OGlobAlignE = ../bin/utility/GlobAlignE.o
+
+# Non TR
+OChromosome  = ../bin/nonltr/Chromosome.o
+OChromosomeOneDigit = ../bin/nonltr/ChromosomeOneDigit.o
+OChromosomeRandom = ../bin/nonltr/ChromosomeRandom.o
+OChromListMaker = ../bin/nonltr/ChromListMaker.o
+OTableBuilder = ../bin/nonltr/TableBuilder.o
+OScorer = ../bin/nonltr/Scorer.o
+ODetectorMaxima = ../bin/nonltr/DetectorMaxima.o
+OChromDetectorMaxima =  ../bin/nonltr/ChromDetectorMaxima.o
+OHMM = ../bin/nonltr/HMM.o
+OScanner = ../bin/nonltr/Scanner.o
+OTrainer = ../bin/nonltr/Trainer.o
+OLocationList = ../bin/nonltr/LocationList.o
+OLocationListCollection = ../bin/nonltr/LocationListCollection.o
+
+OBJS = $(ORed) $(OInvalidInputException) $(OInvalidStateException) $(OFileDoesNotExistException) $(OInvalidOrderOfOperationsException) $(OInvalidOperationException) $(OInvalidScoreException) $(OUtil) $(OLocation) $(OEmptyLocation) $(OChromosome) $(OChromosomeOneDigit) $(OChromosomeRandom) $(OChromListMaker) $(OTableBuilder) $(OScorer) $(ODetectorMaxima) $(OChromDetector) $(OChromDetectorMaxima) $(OHMM) $(OScanner) $(OTrainer) $(OLocationList) $(OLocationListCollection) $(OLCSLen) $(OAffineId) $(OGlobAlignE)
+
+#
+# Target
+#
+
+TRed = ../bin/Red
+
+#
+# Make RepeatsDetector
+#
+
+$(TRed): $(OBJS)
+	$(CXX) -o $(TRed) $(OBJS)
+
+#
+# RepeatsDetector
+#
+
+$(ORed): RepeatsDetector.cpp nonltr/KmerHashTable.h nonltr/KmerHashTable.cpp nonltr/TableBuilder.h nonltr/HMM.h nonltr/Scanner.h nonltr/Trainer.h utility/Util.h
+	$(CXX) $(CXXFLAGS) -c RepeatsDetector.cpp -o $(ORed)
+
+#
+# Exception
+#
+$(OInvalidInputException): exception/InvalidInputException.cpp exception/InvalidInputException.h
+	$(CXX) $(CXXFLAGS) -c exception/InvalidInputException.cpp -o $(OInvalidInputException)
+
+$(OInvalidStateException): exception/InvalidStateException.cpp exception/InvalidStateException.h
+	$(CXX) $(CXXFLAGS) -c exception/InvalidStateException.cpp -o $(OInvalidStateException)
+
+$(OFileDoesNotExistException): exception/FileDoesNotExistException.cpp exception/FileDoesNotExistException.h
+	$(CXX) $(CXXFLAGS) -c exception/FileDoesNotExistException.cpp -o $(OFileDoesNotExistException)
+
+$(OInvalidOrderOfOperationsException): exception/InvalidOrderOfOperationsException.cpp exception/InvalidOrderOfOperationsException.h
+	$(CXX) $(CXXFLAGS) -c exception/InvalidOrderOfOperationsException.cpp -o $(OInvalidOrderOfOperationsException)
+
+$(OInvalidScoreException): exception/InvalidScoreException.cpp exception/InvalidScoreException.h
+	$(CXX) $(CXXFLAGS) -c exception/InvalidScoreException.cpp -o $(OInvalidScoreException)
+
+$(OInvalidOperationException): exception/InvalidOperationException.cpp exception/InvalidOperationException.h
+	$(CXX) $(CXXFLAGS) -c exception/InvalidOperationException.cpp -o $(OInvalidOperationException)
+
+#
+# Utility
+#
+
+$(OUtil): utility/Util.cpp utility/Util.h utility/Location.h exception/FileDoesNotExistException.h
+	$(CXX) $(CXXFLAGS) -c utility/Util.cpp -o $(OUtil)
+
+$(OLocation): utility/Location.cpp utility/Location.h utility/ILocation.h exception/InvalidInputException.h utility/Util.h
+	$(CXX) $(CXXFLAGS) -c utility/Location.cpp -o $(OLocation)
+
+$(OEmptyLocation): utility/EmptyLocation.cpp utility/EmptyLocation.h utility/ILocation.h exception/InvalidOperationException.h
+	$(CXX) $(CXXFLAGS) -c utility/EmptyLocation.cpp -o $(OEmptyLocation)
+
+$(OLCSLen): utility/LCSLen.cpp utility/LCSLen.h
+	$(CXX) $(CXXFLAGS) -c utility/LCSLen.cpp -o $(OLCSLen)
+
+$(OAffineId): utility/AffineId.cpp utility/AffineId.h
+	$(CXX) $(CXXFLAGS) -c utility/AffineId.cpp -o $(OAffineId)
+
+$(OGlobAlignE): utility/GlobAlignE.cpp utility/GlobAlignE.h
+	$(CXX) $(CXXFLAGS) -c utility/GlobAlignE.cpp -o $(OGlobAlignE)
+#
+# Non LTR
+#
+
+$(OChromosome): nonltr/Chromosome.cpp nonltr/Chromosome.h nonltr/IChromosome.h utility/Util.h exception/InvalidInputException.h exception/InvalidOperationException.h
+	$(CXX) $(CXXFLAGS) -c nonltr/Chromosome.cpp -o $(OChromosome)
+
+$(OChromosomeOneDigit): nonltr/ChromosomeOneDigit.cpp nonltr/ChromosomeOneDigit.h nonltr/Chromosome.h exception/InvalidInputException.h
+	$(CXX) $(CXXFLAGS) -c nonltr/ChromosomeOneDigit.cpp	-o $(OChromosomeOneDigit)
+
+$(OChromosomeRandom): nonltr/ChromosomeRandom.cpp nonltr/ChromosomeRandom.h nonltr/IChromosome.h exception/InvalidInputException.h exception/InvalidStateException.h utility/Util.h
+	$(CXX) $(CXXFLAGS) -c nonltr/ChromosomeRandom.cpp -o $(OChromosomeRandom)
+
+$(OTableBuilder): nonltr/TableBuilder.cpp nonltr/TableBuilder.h utility/Util.h nonltr/ChromosomeOneDigit.h nonltr/ITableView.h nonltr/KmerHashTable.h nonltr/KmerHashTable.cpp nonltr/EnrichmentMarkovView.h nonltr/EnrichmentMarkovView.cpp exception/InvalidStateException.h nonltr/ChromListMaker.h nonltr/IChromosome.h
+	$(CXX) $(CXXFLAGS) -c nonltr/TableBuilder.cpp -o $(OTableBuilder)
+
+$(OScorer): nonltr/Scorer.cpp nonltr/Scorer.h nonltr/ChromosomeOneDigit.h utility/Util.h exception/InvalidStateException.h
+	$(CXX) $(CXXFLAGS) -c nonltr/Scorer.cpp -o $(OScorer)
+
+$(ODetectorMaxima): nonltr/DetectorMaxima.cpp nonltr/DetectorMaxima.h utility/ILocation.h exception/InvalidStateException.h
+	$(CXX) $(CXXFLAGS) -c nonltr/DetectorMaxima.cpp -o $(ODetectorMaxima)
+
+$(OChromDetectorMaxima): nonltr/ChromDetectorMaxima.cpp nonltr/ChromDetectorMaxima.h nonltr/DetectorMaxima.h nonltr/ChromosomeOneDigit.h utility/Util.h utility/ILocation.h utility/Location.h
+	$(CXX) $(CXXFLAGS) -c nonltr/ChromDetectorMaxima.cpp -o $(OChromDetectorMaxima)
+
+$(OHMM): nonltr/HMM.cpp nonltr/HMM.h utility/ILocation.h exception/InvalidStateException.h exception/InvalidInputException.h exception/FileDoesNotExistException.h exception/InvalidOperationException.h
+	$(CXX) $(CXXFLAGS) -c nonltr/HMM.cpp -o $(OHMM)
+
+$(OScanner): nonltr/Scanner.cpp nonltr/Scanner.h nonltr/Chromosome.h nonltr/ChromosomeOneDigit.h nonltr/HMM.h nonltr/ITableView.h nonltr/Scorer.h utility/Util.h utility/ILocation.h exception/InvalidInputException.h exception/InvalidStateException.h exception/FileDoesNotExistException.h exception/InvalidOperationException.h
+	$(CXX) $(CXXFLAGS) -c nonltr/Scanner.cpp -o $(OScanner)
+
+$(OTrainer): nonltr/Trainer.cpp nonltr/Trainer.h nonltr/TableBuilder.h nonltr/KmerHashTable.h nonltr/KmerHashTable.cpp nonltr/HMM.h nonltr/ChromDetectorMaxima.h nonltr/Scorer.h nonltr/ChromListMaker.h utility/Util.h nonltr/LocationListCollection.h
+	$(CXX) $(CXXFLAGS) -c nonltr/Trainer.cpp -o $(OTrainer)
+
+$(OChromListMaker): nonltr/ChromListMaker.cpp nonltr/ChromListMaker.h nonltr/Chromosome.h nonltr/ChromosomeOneDigit.h utility/Util.h
+	$(CXX) $(CXXFLAGS) -c nonltr/ChromListMaker.cpp -o $(OChromListMaker)
+
+$(OCluster): nonltr/Cluster.cpp nonltr/Cluster.h utility/Util.h exception/InvalidStateException.h exception/InvalidInputException.h
+	$(CXX) $(CXXFLAGS) -c nonltr/Cluster.cpp -o $(OCluster)
+
+$(OLocationList): nonltr/LocationList.cpp nonltr/LocationList.h utility/ILocation.h utility/Location.h exception/InvalidStateException.h
+	$(CXX) $(CXXFLAGS) -c nonltr/LocationList.cpp -o $(OLocationList)
+
+$(OLocationListCollection): nonltr/LocationListCollection.cpp nonltr/LocationListCollection.h utility/Location.h exception/InvalidStateException.h
+	$(CXX) $(CXXFLAGS) -c nonltr/LocationListCollection.cpp -o $(OLocationListCollection)
+
+
+#
+# Make binary directories
+#
+
+red: $(TRed)
+
+#
+# Make Red
+#
+
+bin:
+	mkdir ../bin
+	mkdir ../bin/exception
+	mkdir ../bin/utility
+	mkdir ../bin/nonltr
+
+#
+# Make clean
+#
+
+clean:
+	rm -f ../bin/*.o ../bin/exception/*.o ../bin/ms/*.o ../bin/nonltr/*.o ../bin/test/*.o ../bin/utility/*.o ../bin/tr/*.o *.o $(TRed)