Merge pull request #13 from jesvedberg/master

Merged with Ancestry_HMM-S and prepared for distribution with Bioconda

src/Makefile

@@ -2,20 +2,35 @@
 
 TCFLAGS = -ltcmalloc 
 ARMAFLAGS = -larmadillo
+CONDAFLAGS = -fvisibility-inlines-hidden -std=c++17 -fmessage-length=0 -march=nocona -mtune=haswell -ftree-vectorize -fPIC -fstack-protector-strong -fno-plt -O2 -ffunction-sections -pipe -isystem ${CONDA_PREFIX}/include
 
 all:
-	$(LINK.cc) -O3 -o ancestry_hmm ancestry_hmm.cpp $(ARMAFLAGS)
-### if tcmalloc is installed, this can be linked using the command below instead
-#	$(LINK.cc) -O3 -o ancestry_hmm ancestry_hmm.cpp $(TCFLAGS) $(ARMAFLAGS)
+	$(CXX) -O3 $(CXXFLAGS) -o ancestry_hmm ancestry_hmm.cpp  $(ARMAFLAGS)
+	$(CXX) -O3 $(CXXFLAGS) -o ahmm-s ahmms.cpp  $(ARMAFLAGS)
+
+conda:
+	$(CXX) -O3 $(CONDAFLAGS) -o ahmm-s ahmms.cpp -L ${CONDA_PREFIX}/lib -I ${CONDA_PREFIX}/include $(ARMAFLAGS)
+	$(CXX) -O3 $(CONDAFLAGS) -o ancestry_hmm ancestry_hmm.cpp -L ${CONDA_PREFIX}/lib -I ${CONDA_PREFIX}/include $(ARMAFLAGS)
+
+ahmms:
+	$(CXX) -O3 $(CXXFLAGS) -o ahmm-s ahmms.cpp $(ARMAFLAGS)
+
+ahmm:
+	$(CXX) -O3 $(CXXFLAGS) -o ancestry_hmm ancestry_hmm.cpp $(ARMAFLAGS)
+
+
+## if you have a local install of google perftools, please add  a TCFlag link. 
+###  $(LINK.cc) -std=c++11 -O3 -o ahmm-s ahmms.cpp $(ARMAFLAGS)  $(TCFLAGS)
 
 ## if you have a local armadillo installation, you will need to provide the directory during compile time and possible also link lblas and lapack 
 ## our recommendation is to use miniconda3 to do the installation
 ##	$ conda install -c conda-forge armadillo
 ## then you will have the appropriate lib and include files in your home directory under subdirectory miniconda3/
 ## so, replace USERNAME with your unix id on the following line and try this
 
-       #$(LINK.cc) -O3 -o ancestry_hmm ancestry_hmm.cpp -L /home/USERNAME/miniconda3/lib/ -I /home/USERNAME/miniconda3/include/ $(ARMAFLAGS)
+##	$(CXX) -std=c++11 -O3 -o ahmm-s ahmms.cpp -L ${CONDA_PREFIX}/lib -I ${CONDA_PREFIX}/include $(ARMAFLAGS)
 
-## if it builds correctly, you will also need to link the library during runtime
+## if it builds correctly, you may also need to link the library during runtime
 ## to do this, add the following line to your ~/.bash_profile or ~/.bashrc
-##     export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/home/USERNAME/miniconda3/lib/
+##     export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:${CONDA_PREFIX}/lib
+

src/ahmms.cpp

@@ -0,0 +1,215 @@
+/*
+
+ copyright: Russ Corbett-Detig
+            rucorbet@ucsc.edu
+
+            Jesper Svedberg
+            jsvedber@ucsc.edu
+
+ This is software distributed under the gnu public license version 3.
+ 
+ */
+
+/// headers
+#include <iostream>
+#include <vector>
+#include <map>
+#include <time.h>
+#include <string>
+#include <fstream>
+#include <algorithm>
+
+// Includes specific for Ancestry_HMM-S
+#include <cmath>  
+#include <cstring>
+#include <utility>
+#include <iomanip>
+#include <cstdlib>
+#include <random> // ++++ REQUIRES C++11 ++++
+
+
+/// linear algebra library is armadillo
+#define ARMA_NO_DEBUG
+#include <armadillo>
+using namespace arma ;
+using namespace std ;
+
+/// our header files in /src directory
+#include "selection_print_usage.h" // JS
+#include "factorial.h"
+#include "nchoosek.h" 
+#include "selection_subsample.h" 
+#include "multichoose.h"
+#include "multipermute.h"
+#include "normalize.h"
+#include "ancestry_pulse.h"
+#include "ploidy_path.h" 
+#include "selection_class.h" // JS
+#include "selection_markov_chain.h"
+#include "read_samples.h" 
+#include "pulses_to_ancestry.h" 
+#include "compute_forward.h"
+#include "compute_backward.h"
+#include "forward_backward.h"
+#include "viterbi.h" 
+#include "transition_information.h"
+#include "exponentiate_matrix.h"
+#include "selection_cmd_line.h"
+#include "create_transition_rates.h"
+#include "selection_read_cmd_line.h"
+#include "evaluate_vertex.h"
+#include "check_vertex.h"
+#include "sort_vertices.h"
+#include "create_pulses.h" 
+#include "create_states.h"
+#include "input_line.h"
+#include "distribute_alleles.h" 
+#include "binomial.h"
+#include "read_emissions.h"
+#include "genotype_emissions.h"
+#include "selection_read_input.h"
+#include "nelder_mead.h"
+#include "golden_search.h"
+#include "bootstrap.h"
+
+// Includes specific for Ancestry_HMM-S
+#include "selection_get_position.h"
+#include "selection_optimize_test_func.h" // Function for testing Nelder-Mead. Remove?
+#include "selection_fwd_iter.h"
+#include "selection_trajectory.h"
+#include "selection_split_vector.h"
+#include "selection_forward.h"
+#include "selection_stochastic_traj.h"
+#include "selection_transition_rates.h"
+
+
+
+
+int main ( int argc, char *argv[] ) {
+
+    /// time tracking
+    clock_t t = clock() ;
+    clock_t total = clock() ;
+
+    /// seed prng
+    srand (t) ;
+
+	// read cmd line 
+	cmd_line options ;
+    cerr << "reading command line" ; t = clock();
+	options.read_cmd_line( argc, argv ) ;
+
+    /// chain objects for each sample
+    vector<markov_chain> markov_chain_information ;
+
+    /// get sample ids and ploidy from input file
+    cerr << "\t\t\t\t" << (double) (clock() - t) << " ms\n" << "reading sample ids and ploidy" ; t = clock();
+    read_samples( markov_chain_information, options.sample_file, options.viterbi ) ;
+
+    /// create states matrix
+    cerr << "\t\t\t" << (double) (clock() - t) << " ms\n" << "creating states matrix" ; t = clock();
+    /// store all possible state space arranged by ploidy and then vector of state counts
+    map<int,vector<vector<int> > > state_list ;
+    /// now create initial state list
+    for ( int m = 0 ; m < markov_chain_information.size() ; m ++ ) {
+        for ( int p = 0 ; p < markov_chain_information[m].sample_ploidy_path.size() ; p ++ ) {
+            create_initial_states( markov_chain_information.at(m).sample_ploidy_path[p].ploidy, options.ancestry_pulses, state_list ) ;
+        }
+    }
+
+	/// read in panels and update matrices
+    cerr << "\t\t\t\t" << (double) (clock() - t) << " ms\n" << "reading data and creating emissions matrices\t" ; t = clock() ;
+    /// store recombination rates and positions
+    vector<int> position ;
+    vector<double> recombination_rate ;
+    vector<string> chromosomes ;
+    int sel_pos ;
+    read_file( options, markov_chain_information, state_list, position, recombination_rate, chromosomes, sel_pos ) ;
+
+
+
+    /// create basic transition information
+    cerr << (double) (clock() - t) << " ms" << endl << "computing transition routes\t\t\t" ; t = clock() ;
+    /// 3d map to look up by ploidy, start state, end state, and then relevant transition information
+    map<int, vector<vector< map< vector<transition_information>, double > > > > transition_matrix_information ;
+    for ( int m = 0 ; m < markov_chain_information.size() ; m ++ ) {
+        for ( int p = 0 ; p < markov_chain_information[m].sample_ploidy_path.size() ; p ++ ) {
+            create_transition_information( markov_chain_information.at(m).sample_ploidy_path[p].ploidy, transition_matrix_information, state_list[markov_chain_information.at(m).sample_ploidy_path[p].ploidy] ) ;
+        }
+    }
+    cerr << endl;
+
+
+    // Below are ahmm-s specific options
+
+    // If using grid search with --grid flag
+    if (options.calc_grid == true) {
+        int p_start = options.grid_pstart;
+        int p_stop = options.grid_pstop;
+        int p_step = options.grid_pstep;
+
+        double s_start = options.grid_sstart;
+        double s_stop = options.grid_sstop;
+
+        if ( options.limit_sel_space == true ) {
+            s_stop = selection_get_max_sel(options.grid_sstart, options.grid_sstop, options.grid_sstep, options.ancestry_pulses[1].proportion, options.ancestry_pulses[1].time, options.ne);
+        }
+        double s_step = options.grid_sstep;
+
+        cerr << "Grid search. Likelihood calculated for values of selection between " << s_start << " and " << s_stop << endl;
+
+        selection_grid(p_start, p_stop, p_step, s_start, s_stop, s_step, markov_chain_information, transition_matrix_information, recombination_rate, position, options, state_list);
+        return 0;
+    }
+
+
+
+    // If testing a single point using --site flag.
+    if (options.test_point == true) {
+        cerr << "Evaluating point: " << options.test_pos << ", " << options.test_sel << endl;
+
+        map <double,vector<double> > sel_trajectories;
+        vector <vector<double> > split_vecs;
+        int testpos;
+        selection point0;
+
+        if (options.is_coord ==  true) {
+            testpos = get_position(options.test_pos, position);
+
+            if (testpos == -1) {
+                cerr << "ERROR: specified site not found on chromosome" << endl;
+                exit(1);
+            }
+
+        }
+        else {
+            testpos = options.test_pos;
+        }
+
+        point0.pos = testpos;
+        point0.sel = 0;
+        selection_evaluate_point_genotypes( point0, markov_chain_information, transition_matrix_information, recombination_rate, position, options, state_list, split_vecs, sel_trajectories ) ;
+
+        selection point1;
+        point1.pos = testpos;
+        point1.sel = options.test_sel;
+        selection_evaluate_point_genotypes( point1, markov_chain_information, transition_matrix_information, recombination_rate, position, options, state_list, split_vecs, sel_trajectories ) ;
+
+        cout << "lnL for a selected site s=" << options.test_sel << " at position " << position[point0.pos] << " is: " << point1.lnl-point0.lnl << endl;
+
+        return 0;
+    }
+
+
+
+    // If using Golden Section Search with --gss flag
+    if (options.run_gss == true) {
+        selection_golden_section(markov_chain_information, transition_matrix_information, recombination_rate, position, options, state_list);
+        return 0;
+    }
+
+
+	return 0 ; 
+}
+
+

src/read_cmd_line.h

@@ -139,7 +139,7 @@ void cmd_line::read_cmd_line ( int argc, char *argv[] ) {
 
         if ( strcmp(argv[i],"--help") == 0 ) {
             print_usage() ;
-            exit(1) ; 
+            exit(0) ; 
         }
 
         if ( strcmp(argv[i],"-g") == 0 ) {

src/selection_class.h

@@ -0,0 +1,21 @@
+#ifndef __SELECTION_CLASS_H
+#define __SELECTION_CLASS_H
+
+class selection {
+public:
+    int pos;
+    double sel;
+    double lnl;
+
+    /// sort pulses by time
+    friend bool operator < ( const selection &a, const selection &b ) {
+        return a.lnl < b.lnl ;
+    }
+} ;
+
+ostream& operator<< (ostream &out, selection const& point) {
+    out << "Selection point. pos:" << point.pos << "  sel:" << setprecision(15) << point.sel << "  lnL: "  << point.lnl;
+    return out;
+}
+
+#endif