ngs_server 0.1 → 0.2

data/ext/vcftools/cpp/.svn/text-base/vcftools.cpp.svn-base

@@ -0,0 +1,107 @@
+/*
+ * vcftools.cpp
+ *
+ *  Created on: Aug 19, 2009
+ *      Author: Adam Auton
+ *      ($Revision: 249 $)
+ */
+#include "vcftools.h"
+
+ofstream LOG;
+
+int main(int argc, char *argv[])
+{
+	time_t start,end;
+	time(&start);
+
+	// The following turns off sync between C and C++ streams.
+	// Apparently it's faster to turn sync off, and as I don't use C streams, it's okay to turn off.
+	ios_base::sync_with_stdio(false);
+
+	parameters params(argc, argv);
+	params.print_help();
+	params.read_parameters();
+
+	LOG.open((params.output_prefix + ".log").c_str());
+
+	printLOG("\nVCFtools - " + VCFTOOLS_VERSION + "\n");
+	printLOG("(C) Adam Auton 2009\n\n");
+
+	params.print_params();
+
+	vcf_file vcf(params.vcf_filename, params.vcf_compressed, params.chr_to_keep, params.chr_to_exclude, params.force_write_index);
+
+	// Apply various filters as required.
+	vcf.apply_filters(params);
+
+	unsigned int N_indv = vcf.N_kept_individuals();
+	unsigned int N_sites = vcf.N_kept_sites();
+	printLOG("After filtering, kept " + int2str(N_indv) + " out of " + int2str(vcf.N_indv) + " Individuals\n");
+	printLOG("After filtering, kept " + int2str(N_sites) + " out of a possible " + int2str(vcf.N_entries) + " Sites\n");
+	if (N_sites == 0)
+		error("No data left for analysis!");
+
+	if (params.diff_file != "")
+	{	// Merge files - cannot be run with other output options.
+		vcf_file vcf_diff(params.diff_file, params.diff_file_compressed, params.chr_to_keep, params.chr_to_exclude, params.force_write_index);
+		vcf_diff.apply_filters(params);	// Apply various filters as required.
+		vcf.output_indv_in_files(params.output_prefix, vcf_diff);
+		vcf.output_sites_in_files(params.output_prefix, vcf_diff);
+
+		if (params.diff_site_discordance == true) vcf.output_discordance_by_site(params.output_prefix, vcf_diff);
+		if (params.diff_discordance_matrix == true) vcf.output_discordance_matrix(params.output_prefix, vcf_diff);
+		if (params.diff_indv_discordance == true) vcf.output_discordance_by_indv(params.output_prefix, vcf_diff);
+		if (params.diff_switch_error == true) vcf.output_switch_error(params.output_prefix, vcf_diff);
+	}
+
+	vcf.output_INFO_for_each_site(params.output_prefix, params.INFO_to_extract);
+	vcf.output_FORMAT_information(params.output_prefix, params.FORMAT_id_to_extract);
+	if (params.output_indv_depth == true) vcf.output_individuals_by_mean_depth(params.output_prefix);
+	if (params.output_geno_depth == true) vcf.output_genotype_depth(params.output_prefix);
+	if (params.output_site_depth == true) vcf.output_site_depth(params.output_prefix, false);
+	if (params.output_site_mean_depth == true) vcf.output_site_depth(params.output_prefix, true);
+	if (params.output_freq == true) vcf.output_frequency(params.output_prefix, false, params.suppress_allele_output);
+	if (params.output_counts == true) vcf.output_frequency(params.output_prefix, true, params.suppress_allele_output);
+	if (params.plink_output == true) vcf.output_as_plink(params.output_prefix);
+	if (params.plink_tped_output == true) vcf.output_as_plink_tped(params.output_prefix);
+	if (params.output_HWE == true) vcf.output_hwe(params.output_prefix);
+	if (params.output_SNP_density_bin_size > 0) vcf.output_SNP_density(params.output_prefix, params.output_SNP_density_bin_size);
+	if (params.output_missingness == true) vcf.output_missingness(params.output_prefix);
+	if (params.output_geno_rsq == true) vcf.output_genotype_r2(params.output_prefix, params.ld_snp_window_size, params.ld_bp_window_size, params.min_r2);
+	if (params.output_interchromosomal_rsq == true) vcf.output_interchromosomal_genotype_r2(params.output_prefix, params.min_r2);
+	if (params.output_hap_rsq == true) vcf.output_haplotype_r2(params.output_prefix, params.ld_snp_window_size, params.ld_bp_window_size, params.min_r2);
+	if (params.output_het == true) vcf.output_het(params.output_prefix);
+	if (params.output_site_quality == true) vcf.output_site_quality(params.output_prefix);
+	if (params.output_012_matrix == true) vcf.output_as_012_matrix(params.output_prefix);
+	if (params.output_as_IMPUTE == true) vcf.output_as_IMPUTE(params.output_prefix);
+	if (params.output_BEAGLE_genotype_likelihoods == true) vcf.output_BEAGLE_genotype_likelihoods(params.output_prefix);
+	if (params.output_as_ldhat_unphased == true) vcf.output_as_LDhat_unphased(params.output_prefix, params.chr_to_keep);
+	if (params.output_as_ldhat_phased == true) vcf.output_as_LDhat_phased(params.output_prefix, params.chr_to_keep);
+	if (params.output_singletons == true) vcf.output_singletons(params.output_prefix);
+	if (params.output_site_pi == true) vcf.output_per_site_nucleotide_diversity(params.output_prefix);
+	if (params.pi_window_size > 0) vcf.output_windowed_nucleotide_diversity(params.output_prefix, params.pi_window_size);
+	if (params.output_Tajima_D_bin_size > 0) vcf.output_Tajima_D(params.output_prefix, params.output_Tajima_D_bin_size);
+	if (params.output_TsTv_bin_size > 0) vcf.output_TsTv(params.output_prefix, params.output_TsTv_bin_size);
+	if (params.output_TsTv_by_count) vcf.output_TsTv_by_count(params.output_prefix);
+	if (params.output_TsTv_by_qual) vcf.output_TsTv_by_quality(params.output_prefix);
+	if (params.recode == true) vcf.print(params.output_prefix, params.recode_INFO_to_keep, params.recode_all_INFO);
+	if (params.output_filter_summary == true) vcf.output_FILTER_summary(params.output_prefix);
+	if (params.output_filtered_sites == true) vcf.output_kept_and_removed_sites(params.output_prefix);
+	if (params.output_LROH == true) vcf.output_LROH(params.output_prefix);
+	if (params.output_relatedness == true) vcf.output_indv_relatedness(params.output_prefix);
+	if (params.output_PCA == true) vcf.output_PCA(params.output_prefix, !params.PCA_no_normalisation, params.output_N_PCA_SNP_loadings);
+	if (params.fst_populations.size() > 0) vcf.output_fst_version_2(params.output_prefix, params.fst_populations);
+
+	if (params.fst_file != "")
+	{
+		vcf_file vcf_fst(params.fst_file, params.fst_file_compressed, params.chr_to_keep);
+		vcf_fst.apply_filters(params);	// Apply various filters as required.
+		vcf.output_fst(params.output_prefix, vcf_fst);
+	}
+
+	time(&end);
+	double running_time = difftime(end,start);
+	printLOG("Run Time = " + dbl2str_fixed(running_time, 2) + " seconds\n");
+	LOG.close();
+	return 0;
+}

data/ext/vcftools/cpp/.svn/text-base/vcftools.h.svn-base

@@ -0,0 +1,25 @@
+/*
+ * vcftools.h
+ *
+ *  Created on: Aug 19, 2009
+ *      Author: Adam Auton
+ *      ($Revision: 122 $)
+ */
+
+#ifndef VCFTOOLS_H_
+#define VCFTOOLS_H_
+
+#include <cstring>
+#include <ctime>
+#include <iostream>
+#include <limits>
+#include <string>
+#include <vector>
+
+#include "output_log.h"
+#include "parameters.h"
+#include "vcf_file.h"
+
+using namespace std;
+
+#endif/* VCFTOOLS_H_ */

data/ext/vcftools/cpp/Makefile

@@ -0,0 +1,46 @@
+# Make file for vcftools
+# Author: Adam Auton
+# ($Revision: 230 $)
+
+# Compiler
+CPP = g++
+# Output executable
+EXECUTABLE = vcftools
+# Flag used to turn on compilation of PCA routines
+ifndef VCFTOOLS_PCA
+	VCFTOOLS_PCA = 0
+endif
+# Compiler flags
+CPPFLAGS = -O2 -Wall -Wextra
+#CPPFLAGS = -g 
+# Included libraries (zlib)
+LIB = -lz 
+#LIB = -lz -I/opt/local/include/ -L/opt/local/lib/
+SOURCES = vcftools.cpp vcf_file.cpp vcf_entry.cpp \
+			vcf_entry_getters.cpp vcf_entry_setters.cpp \
+			vcf_file_filters.cpp vcf_file_output.cpp \
+			vcf_file_format_convert.cpp \
+			vcf_file_diff.cpp parameters.cpp \
+			vcf_file_index.cpp \
+			output_log.cpp
+
+ifeq ($(VCFTOOLS_PCA), 1)
+	# Define flag for PCA routine compilation
+	CPPFLAGS += -DVCFTOOLS_PCA
+	# Add LAPACK library
+	LIB += -llapack	
+	# Add PCA source code
+	SOURCES += dgeev.cpp
+endif
+
+all: vcftools
+
+vcftools: $(SOURCES)
+	$(CPP) $(CPPFLAGS) $(SOURCES) -o $@ $(LIB)
+ifdef BINDIR
+	cp $(CURDIR)/$@ $(BINDIR)/$@
+endif
+
+clean:
+	@rm -f vcftools
+	@rm -f $(BINDIR)/vcftools

data/ext/vcftools/cpp/dgeev.cpp

@@ -0,0 +1,146 @@
+/*
+ * dgeev.cpp
+ *
+ *	Created on: 20 Apr 2011
+ *			Author: auton
+ * ($Revision: 1 $)
+ */
+
+#include "dgeev.h"
+
+void dgeev_sort(double *Er, double *Ei, int N)
+{
+	double temp, *E2;
+	int i, j;
+
+	E2 = new double[N];
+	for (i=0; i<N; i++)
+		E2[i] = Er[i]*Er[i]+Ei[i]*Ei[i];
+
+	for (j=0; j<N; j++)
+		for (i=0; i<N-1; i++)
+			if (fabs(E2[i])<fabs(E2[i+1]))
+			{
+				temp = E2[i]; E2[i] = E2[i+1]; E2[i+1] = temp;
+				temp = Er[i]; Er[i] = Er[i+1]; Er[i+1] = temp;
+				temp = Ei[i]; Ei[i] = Ei[i+1]; Ei[i+1] = temp;
+			}
+
+	delete [] E2;
+}
+
+
+void dgeev_sort(double *Er, double *Ei, double **Evecs, int N)
+{
+	double temp, *E2;
+	int i, j, k;
+
+	E2 = new double[N];
+	for (i=0; i<N; i++)
+		E2[i] = Er[i]*Er[i]+Ei[i]*Ei[i];
+
+	for (j=0; j<N; j++)
+		for (i=0; i<N-1; i++)
+			if (fabs(E2[i])<fabs(E2[i+1]))
+			{
+				temp = E2[i]; E2[i] = E2[i+1]; E2[i+1] = temp;
+				temp = Er[i]; Er[i] = Er[i+1]; Er[i+1] = temp;
+				temp = Ei[i]; Ei[i] = Ei[i+1]; Ei[i+1] = temp;
+
+				for (k=0; k<N; k++)
+				{
+					temp = Evecs[k][i];
+					Evecs[k][i] = Evecs[k][i+1];
+					Evecs[k][i+1] = temp;
+				}
+			}
+
+	delete [] E2;
+}
+
+
+double* dgeev_ctof(double **in, int rows, int cols)
+{
+	double *out;
+	int i, j;
+
+	out = new double[rows*cols];
+	for (i=0; i<rows; i++)
+		for (j=0; j<cols; j++)
+			out[i+j*cols] = in[i][j];
+	return(out);
+}
+
+
+void dgeev_ftoc(double *in, double **out, int rows, int cols)
+{
+	int i, j;
+
+	for (i=0; i<rows; i++)
+		for (j=0; j<cols; j++)
+			out[i][j] = in[i+j*cols];
+}
+
+
+void dgeev(double **H, int n, double *Er, double *Ei)
+{
+	char jobvl, jobvr;
+	int lda, ldvl, ldvr, lwork, info;
+	double *a, *vl, *vr, *work;
+
+	jobvl = 'N'; // V/N to calculate/not calculate the left eigenvectors of the matrix H.
+	jobvr = 'N'; // As above, but for the right eigenvectors.
+
+	lda = n; // The leading dimension of the matrix a.
+	a = dgeev_ctof(H, n, lda); // Convert the matrix H from double pointer C form to single pointer Fortran form.
+
+	/* Whether we want them or not, we need to define the matrices
+		 for the eigenvectors, and give their leading dimensions.
+		 We also create a vector for work space. */
+
+	ldvl = n;
+	vl = new double[n*n];
+	ldvr = n;
+	vr = new double[n*n];
+	work = new double[4*n];
+	lwork = 4*n;
+
+	dgeev_(&jobvl, &jobvr, &n, a, &lda, Er, Ei, vl, &ldvl, vr, &ldvr, work, &lwork, &info);
+
+	dgeev_sort(Er, Ei, n); //Sort the results by eigenvalue in decreasing magnitude.
+
+	delete [] a;
+	delete [] vl;
+	delete [] vr;
+	delete [] work;
+}
+
+
+void dgeev(double **H, int n, double *Er, double *Ei, double **Evecs)
+{
+	char jobvl, jobvr;
+	int lda,	ldvl, ldvr, lwork, info;
+	double *a, *vl, *vr, *work;
+
+	jobvl = 'N';
+	jobvr = 'V';
+	lda = n;
+	a = dgeev_ctof(H, n, lda);
+
+	ldvl = n;
+	vl = new double[n*n];
+	ldvr = n;
+	vr = new double[n*n];
+	work = new double[4*n];
+	lwork = 4*n;
+
+	dgeev_(&jobvl, &jobvr, &n, a, &lda, Er, Ei, vl, &ldvl, vr, &ldvr, work, &lwork, &info);
+
+	dgeev_ftoc(vr, Evecs, n, ldvr);
+	dgeev_sort(Er, Ei, Evecs, n);
+
+	delete [] a;
+	delete [] vl;
+	delete [] vr;
+	delete [] work;
+}

data/ext/vcftools/cpp/dgeev.h

@@ -0,0 +1,43 @@
+/*
+ * dgeev.h
+ * ($Revision: 1 $)
+  This file contains a wrapper around the LAPACK dgeev routine,
+  used to calculate eigenvalues and associated eigenvectors
+  for a square asymmetric matrix H. Since the matrix is asymmetrix,
+  both real and imaginary components of the eigenvalues are returned.
+  For real, symmetric matricies, the eigenvalues will be real, and
+  hence the complex terms equal to zero.
+
+  There are two function calls defined in this header, of the
+  forms
+
+    void dgeev(double **H, int n, double *Er, double *Ei)
+    void dgeev(double **H, int n, double *Er, double *Ei, double **Evecs)
+
+    H: the n by n matrix that we are solving.
+    n: the order of the square matrix H.
+    Er: an n-element array to hold the real parts of the eigenvalues.
+    Ei: an n-element array to hold the imaginary parts of the eigenvalues.
+    Evecs: an n by n matrix to hold the eigenvectors.
+*/
+
+#ifndef DGEEV_H_
+#define DGEEV_H_
+
+#include <cmath>
+
+void dgeev(double **H, int n, double *Er, double *Ei);
+void dgeev(double **H, int n, double *Er, double *Ei, double **Evecs);
+
+double *dgeev_ctof(double **in, int rows, int cols);
+void dgeev_ftoc(double *in, double **out, int rows, int cols);
+void dgeev_sort(double *Er, double *Ei, int N);
+void dgeev_sort(double *Er, double *Ei, double **Evecs, int N);
+
+ 
+extern "C" void dgeev_(char *jobvl, char *jobvr, int *n, double *a,
+		       int *lda, double *wr, double *wi, double *vl,
+		       int *ldvl, double *vr, int *ldvr,
+		       double *work, int *lwork, int *info);
+
+#endif

data/ext/vcftools/cpp/output_log.cpp

@@ -0,0 +1,79 @@
+/*
+ * log.cpp
+ *
+ *  Created on: Nov 11, 2009
+ *      Author: Adam Auton
+ *      ($Revision: 66 $)
+ */
+
+#include "output_log.h"
+
+void printLOG(string s)
+{
+	LOG << s; LOG.flush();
+	cout << s; cout.flush();
+}
+
+void error(string err_msg, int error_code)
+{
+	printLOG("Error:" + err_msg + "\n");
+	exit(error_code);
+}
+
+
+void error(string err_msg, double value1, double value2, int error_code)
+{
+	printLOG("Error:" + err_msg + "\n");
+	stringstream ss;
+	ss << "Value1=" << value1 << " Value2=" << value2 << endl;
+	printLOG(ss.str());
+	exit(error_code);
+}
+
+void warning(string err_msg)
+{
+	printLOG(err_msg + "\n");
+}
+
+void one_off_warning(string err_msg)
+{
+	static set<string> previous_warnings;
+	if (previous_warnings.find(err_msg) == previous_warnings.end())
+	{
+		printLOG(err_msg + "\n");
+		previous_warnings.insert(err_msg);
+	}
+}
+
+string int2str(int n)
+{
+  std::ostringstream s2( std::stringstream::out );
+  s2 << n;
+  return s2.str();
+}
+
+string longint2str(long int n)
+{
+  std::ostringstream s2( std::stringstream::out );
+  s2 << n;
+  return s2.str();
+}
+
+string dbl2str(double n, int prc)
+{
+  std::ostringstream s2;
+  if ( prc > 0 )
+    s2.precision(prc);
+  s2 << n;
+  return s2.str();
+}
+
+string dbl2str_fixed(double n, int prc)
+{
+  std::ostringstream s2;
+  s2 << setiosflags( ios::fixed );
+  if ( prc > 0 )
+    s2.precision(prc);
+  s2 << n;
+  return s2.str();
+}

data/ext/vcftools/cpp/output_log.h

@@ -0,0 +1,34 @@
+/*
+ * log.h
+ *
+ *  Created on: Nov 11, 2009
+ *      Author: Adam Auton
+ *      ($Revision: 91 $)
+ */
+
+#ifndef LOG_H_
+#define LOG_H_
+
+#include <cstdio>
+#include <cstdlib>
+#include <fstream>
+#include <iomanip>
+#include <iostream>
+#include <set>
+#include <sstream>
+
+using namespace std;
+
+extern ofstream LOG;
+
+void printLOG(string s);
+void error(string err_msg, int error_code=0);
+void error(string err_msg, double value1, double value2, int error_code=0);
+void warning(string err_msg);
+void one_off_warning(string err_msg);
+string int2str(int n);
+string longint2str(long int n);
+string dbl2str(double n, int prc);
+string dbl2str_fixed(double n, int prc);
+
+#endif /* LOG_H_ */