adding statistical tests (FET, CHI2), need a probable rewamp

cpp/utils.cpp

@@ -5,14 +5,21 @@
 #include <sstream>
 #include <stdexcept>
 #include <cstring>
+#include <algorithm>
+#include <cmath>
 #include <sys/time.h>
 #include <sys/types.h>
 
-
+#include <boost/math/distributions/chi_squared.hpp>
+#include <boost/math/distributions/hypergeometric.hpp>
 
 #include "utils.h"
 using namespace std;
-
+using namespace boost::math;
+using boost::math::chi_squared;
+using boost::math::quantile;
+using boost::math::complement;
+using boost::math::cdf;
 
 // test if a file is readable
 // return value : true if readable ; false if not
@@ -80,6 +87,9 @@ string char_to_string(char incoming)
 
 vector<string> parseOnSep( string inc , string sep )
 {
+	// cerr << "Entering ParseOnSep function" << endl;
+	// cerr << "\tIncoming string : " << inc << " ; separator : " << sep << endl;
+
 	vector<string> ret;
 	istringstream issInc (inc);
 	string mot;
@@ -116,16 +126,48 @@ string strip( string inc )
 	return inc;
 }
 
+double chisquare( vector<double> toTest , vector<double> all )
+{
+	boost::math::chi_squared chi(1);
+	double a1 = toTest[0] ;
+	double a2 = toTest[1];
+	double b1 = all[0];
+	double b2 = all[1];;
 
 
+	double s = a1 + a2 + b1 + b2;
+	double K = s * (a1 * b2 - a2 * b1) * (a1 * b2 - a2 * b1) / (a1 + a2) / (b1 + b2) / (a1 + b1) / (a2 + b2);
+	double P = boost::math::cdf(chi, K);
 
+	return P;
+}
 
 
-
-
-
-
-
+double fisher_test(vector<double> toTest, vector<double> control )
+{
+	double a = toTest[0];
+	double b = toTest[1];
+	double c = control[0];
+	double d = control[1];
+
+	double N = a + b + c + d;
+	double r = a + c;
+	double n = c + d;
+	double max_for_k = min(r, n);
+	double min_for_k = (double)max(0, int(r + n - N));
+	hypergeometric_distribution<> hgd(r, n, N);
+	double cutoff = pdf(hgd, c);
+	double tmp_p = 0.0;
+	for(int k = min_for_k;k < max_for_k + 1;k++)
+	{
+		double p = pdf(hgd, k);
+		if(p <= cutoff)
+		{
+			tmp_p += p;
+		}
+	}
+	return tmp_p;
+}
 
 
 char checkBase(char incoming )

cpp/utils.h

@@ -15,5 +15,9 @@ std::string char_to_string(char);
 std::vector<std::string> parseOnSep( std::string , std::string );
 char string_to_char( std::string );
 std::string strip( std::string );
+double chisquare( std::vector<double>, std::vector<double> );
+double fisher_test(std::vector<double> , std::vector<double> );
+
+
 
 #endif