void Consolidator::performConsolidation(ErrorCalculator& eCalculator, int gap,int min_snp,float min_cm,int extendSnp)
{
//cout<<"in consolidator extendsnp"<<extendSnp<<endl;
int consolidations = 0, removed = 0;
for(int i=0;i<person_count;++i)//for each person
{
for(int j=i;j<person_count;++j)//compare with each other person
{
int temp1=-1,temp2=-1;
for(int l=0;l<m_matches[i][j].size();++l)//for each match
{
temp1= m_matches[i][j][l].start;//cout<<"temp1 before = "<<temp1<<endl;
temp2= m_matches[i][j][l].end;//cout<<"temp2 before = "<<temp2<<endl;
if(temp2==-1||temp1==-1){continue;}
for(int k=l+1;k<m_matches[i][j].size();++k) //for each other match
{
if((m_matches[i][j][k].start-temp2-1)<=gap)
{
++consolidations;
temp2=m_matches[i][j][k].end;
m_matches[i][j][l].end=temp2;
m_matches[i][j][k].end=-1;
}
else break;
}
//this may be what is causing our initial drops to never show up...
if( ( (temp2-temp1)<min_snp) || ( (eCalculator.getCMDistance(temp2)-eCalculator.getCMDistance(temp1))<min_cm) )
{
++removed;
// m_matches[i][j][l].end=m_matches[i][j][l].start=-1;
}
}
}
}
/*std::string str = " \n Number of Consolidations: " +
NumberToString( consolidations );
str = str + " \n Number of matches removed due to initial length: "
+ NumberToString( removed );*/
/*new*/
global_initial = removed;
consolidated_str = "Number of Consolidations: " + NumberToString( consolidations );
initial_drop_str = "Number of matches removed due to initial length: " + NumberToString( removed );
/*wen*/
// eCalculator.log( str );
}
void Consolidator::performConsolidation(ErrorCalculator& eCalculator, int gap,int min_snp,float min_cm)
{
int consolidations = 0, removed = 0;
for(int i=0;i<person_count;++i)//for each person
{
for(int j=i;j<person_count;++j)//compare with each other person
{
int temp1=-1,temp2=-1;
for(int l=0;l<m_matches[i][j].size();++l)//for each match
{
temp1= m_matches[i][j][l].start;
temp2= m_matches[i][j][l].end;
if(temp2==-1||temp1==-1){continue;}
for(int k=l+1;k<m_matches[i][j].size();++k) //for each other match
{
if((m_matches[i][j][k].start-temp2-1)<=gap)
{
++consolidations;
temp2=m_matches[i][j][k].end;
m_matches[i][j][l].end=temp2;
m_matches[i][j][k].end=-1;
}
else break;
}
if( ( (temp2-temp1)<min_snp) || ( (eCalculator.getCMDistance(temp2)-eCalculator.getCMDistance(temp1))<min_cm) )
{
++removed;
}
}
}
}
/*new*/
global_initial = removed;
consolidated_str = "Number of Consolidations: " + NumberToString( consolidations );
initial_drop_str = "Number of matches removed due to initial length: " + NumberToString( removed );
/*wen*/
}
void Consolidator::readMatches(string path,int pers_count, ErrorCalculator& eCalculator, int trueSNP, float trueCM )
{
person_count=pers_count;
if(pers_count<=0)
{
std::cerr<<"wrong BSID file, check it, reading ped file failed"<<std::endl;
return;
}
try
{
person_count=pers_count;
m_matches.resize(pers_count+1);
m_trueMatches.resize( pers_count + 1 );
for(int i=0;i<pers_count;++i)
{
m_matches[i].resize(pers_count+1);
m_trueMatches[i].resize( pers_count + 1 );
}
unsigned int pid[2];
unsigned int sid[2];
unsigned int dif,hom[2];
ifstream file_bmatch(path.c_str(),ios::binary);
if( !file_bmatch )
{
cerr<<"unable to open the bmatch file, exiting the program" << endl;
exit( -1 );
}
while ( !file_bmatch.eof())
{
pid[0] = -1;
file_bmatch.read( (char*) &pid[0] , sizeof( unsigned int ) );
if ( pid[0] == -1 ) continue;
file_bmatch.read( (char*) &pid[1] , sizeof( unsigned int ) );
file_bmatch.read( (char*) &sid[0] , sizeof( unsigned int ) );
file_bmatch.read( (char*) &sid[1] , sizeof( unsigned int ) );
file_bmatch.read( (char*) &dif , sizeof( int ) );
file_bmatch.read( (char*) &hom[0] , sizeof( bool ) );
file_bmatch.read( (char*) &hom[1] , sizeof( bool ) );
if(pid[0]>=pers_count||pid[1]>=pers_count)
{
cerr<<"problem with bsid file, check it please"<<endl;
return;
}
SNP snp;
snp.start=sid[0];
snp.end=sid[1];
if(pid[0]<=pid[1])
m_matches[(pid[0])][(pid[1])].push_back(snp);
else
m_matches[(pid[1])][(pid[0])].push_back(snp);
if( ( eCalculator.getCMDistance( sid[ 1 ] ) -
eCalculator.getCMDistance( sid[ 0 ] ) ) >= trueCM &&
( sid[ 1 ] - sid[ 0 ] ) >= trueSNP && pid[0] != pid[1] )
{
if(pid[0]<=pid[1])
m_trueMatches[(pid[0])][(pid[1])].push_back(snp);
else
m_trueMatches[(pid[1])][(pid[0])].push_back(snp);
}
}
file_bmatch.close();
}
catch(exception &e)
{
cerr<<"Error:"<<e.what()<<endl;
exit( -1 );
}
}
void Consolidator::findTrueSimplePctErrors( ErrorCalculator &e_obj, float PIElength, bool holdOut,int window, float ma_threshold, float empirical_ma_threshold )
{
for(int i=0;i<person_count;i++)
{
for(int j=i;j<person_count;j++)
{
for(int l=0;l<m_trueMatches[i][j].size();l++)
{
if(m_trueMatches[i][j][l].end==-1)
{
continue;
}
//-------------------------------------------------------------------------------------------------
int t1 = m_trueMatches[ i ][ j ][ l ].start +
( m_trueMatches[ i ][ j ][ l ].end -
m_trueMatches[ i ][ j ][ l ].start ) * 0.25;
int t2 = m_trueMatches[ i ][ j ][ l ].end -
( m_trueMatches[ i ][ j ][ l ].end -
m_trueMatches[ i ][ j ][ l ].start ) * 0.25;
/*What do these two functions do? Is this necessary for being able to find errors, or
is it only useful for MA calculations?*/
vector<vector<int> > trueErrors=e_obj.checkErrors( i, j, t1, t2);
vector<int>finalTrueErrors=e_obj.getFinalErrors( trueErrors );
/*x*/
//This section handles finding the maximum moving averages amongst trulyIBD segments
std::vector<float> av;
float current_max;
if(empirical_ma_threshold < 0.0){
av = e_obj.getTrueMovingAverages(finalTrueErrors,t1,t2,window);
current_max = av[0];
for(int q = 1; q < av.size(); q++){
if(av[q] > current_max){
current_max = av[q];
}
}
e_obj.addMaxAverage(current_max);
}
//------------------------------------------------------------------------------
int temp1 = m_trueMatches[i][j][l].start;
int temp2 = m_trueMatches[i][j][l].end;
float startCM = e_obj.getCMDistance( temp1 );
float endCM = e_obj.getCMDistance( temp2 );
float mid1CM = startCM + ( endCM - startCM ) / 2 - PIElength / 2;
float mid2CM = startCM + ( endCM - startCM ) / 2 + PIElength / 2;
while( e_obj.getCMDistance( temp1 ) <= mid1CM || e_obj.getCMDistance( temp2 ) >=mid2CM )
{
if( e_obj.getCMDistance( temp1 ) <= mid1CM )
{
++temp1;
}
if( e_obj.getCMDistance( temp2 ) >=mid2CM )
{
--temp2;
}
}
/*Here they are again.
*/
vector<vector<int> > errors=e_obj.checkErrors( i, j, temp1, temp2);
vector<int>finalErrors=e_obj.getFinalErrors( errors );
// float per_err = e_obj.getThreshold(finalErrors,temp1, temp2, 0 );
float per_err = e_obj.getThreshold(finalErrors,temp1,temp2); //overload!
m_errors.push_back( per_err );
/*x*/
}//end for(l)
}//end for(j)
}//end for(i)
//this section actually handles the sorting of the max averages, and the setting of the user supplied percentile.
vector<float>maxes;
float cutoff = empirical_ma_threshold; //assume the user wanted to supply a value. This value will be overwritten shortly if they did not.
if(empirical_ma_threshold < 0.0){
maxes = e_obj.getMaxAverages();
std::sort(maxes.begin(),maxes.end());
e_obj.setMaxAverage(maxes);
cutoff = e_obj.getXthPercentile(ma_threshold); //<-make that an actual user input value
}
e_obj.setCutoff(cutoff);//set the actual threshold to be used when calculating MA in all other SH
if(empirical_ma_threshold < 0.0){
ma_thresh_str = "User supplied ma-threshold is: " + NumberToString(ma_threshold);
emp_ma_thresh_str = "Moving Averages will be tested usign the empirical threshold: " + NumberToString(cutoff);
} else {
emp_ma_thresh_str = "Moving Averages will be tested usign the empirical threshold: " + NumberToString(cutoff);
}
//----------------------------------------
std::sort( m_errors.begin(), m_errors.end() );
std::sort( m_holdOutErrors.begin(), m_holdOutErrors.end() );
ibg_str = "No of segments deemed to be IBD for finding empirical error threshold "
+ NumberToString( m_errors.size() );
}//end ftspe
void Consolidator::readMatches(std::string path,int pers_count, ErrorCalculator& eCalculator, int trueSNP, float trueCM, int snipExtend,std::string pedFile )//path->BMATCHFILE file
{
person_count=pers_count;
if(pers_count<=0)
{
std::cerr<<"wrong BSID file, check it, reading ped file failed"<<std::endl;
return;
}
try
{
person_count=pers_count;
m_matches.resize(pers_count+1);
m_trueMatches.resize( pers_count + 1 );
for(int i=0;i<pers_count;++i)
{
m_matches[i].resize(pers_count+1);
m_trueMatches[i].resize( pers_count + 1 );
}
unsigned int pid[2];
unsigned int sid[2];
unsigned int dif,hom[2];
std::ifstream file_bmatch(path.c_str(),std::ios::binary);
if( !file_bmatch )
{
std::cerr<<"unable to open the bmatch file, exiting the program" << std::endl;
std::cout<<"we found the bmatch file"<<std::endl;
exit( -1 );
}
//unsigned long long counttt=0;
while ( !file_bmatch.eof())
{
//counttt++;
pid[0] = -1;
file_bmatch.read( (char*) &pid[0] , sizeof( unsigned int ) );
if ( pid[0] == -1 ) continue;
file_bmatch.read( (char*) &pid[1] , sizeof( unsigned int ) );
file_bmatch.read( (char*) &sid[0] , sizeof( unsigned int ) );
file_bmatch.read( (char*) &sid[1] , sizeof( unsigned int ) );
file_bmatch.read( (char*) &dif , sizeof( int ) );
file_bmatch.read( (char*) &hom[0] , sizeof( bool ) );
file_bmatch.read( (char*) &hom[1] , sizeof( bool ) );
if(pid[0]>=pers_count||pid[1]>=pers_count)
{
std::cerr<<"problem with bsid file, check it please"<<std::endl;
return;
}
SNP_lrf snp;
snp.start=sid[0];
snp.end=sid[1];
//
std::cout<<" pid[0]= "<<pid[0];
std::cout<<"\tpid[1]= "<<pid[1];
std::cout<<"\tsid[0]= "<<sid[0];
std::cout<<"\tsid[1]= "<<sid[1];
std::cout<<"\tdif= "<<dif;
std::cout<<"\thom[0]="<<hom[0];
std::cout<<"\thom[1]= "<<hom[1];
std::cout<<"\tsnp.start = "<<sid[0];
std::cout<<"\tsnp.end=" <<sid[1];
std::cout<<std::endl;
//
if(pid[0]<=pid[1])
m_matches[(pid[0])][(pid[1])].push_back(snp);
else
m_matches[(pid[1])][(pid[0])].push_back(snp);
if( ( eCalculator.getCMDistance( sid[ 1 ] ) -
eCalculator.getCMDistance( sid[ 0 ] ) ) >= trueCM &&
( sid[ 1 ] - sid[ 0 ] ) >= trueSNP && pid[0] != pid[1] )
{
if(pid[0]<=pid[1])
m_trueMatches[(pid[0])][(pid[1])].push_back(snp);
else
m_trueMatches[(pid[1])][(pid[0])].push_back(snp);
}
}
// std::cout<<m_matches.size()<<std::endl;
file_bmatch.close();
}
catch(std::exception &e)
{
std::cerr<<"Error:"<<e.what()<<std::endl;
exit( -1 );
}
}