void gcta::read_phen(string phen_file, vector<string> &phen_ID, vector<string> &phen_buf, int mphen)
{
// Read phenotype data
ifstream in_phen(phen_file.c_str());
if(!in_phen) throw("Error: can not open the file ["+phen_file+"] to read.");
int i=0;
vector<string> fid, pid, vs_buf;
string str_buf, fid_buf, pid_buf;
phen_ID.clear();
phen_buf.clear();
cout<<"Reading phenotypes from ["+phen_file+"]."<<endl;
getline(in_phen, str_buf);
int phen_num=StrFunc::split_string(str_buf, vs_buf)-2;
if(phen_num<=0) throw("Error: no phenotype data is found.");
if(phen_num>1) cout<<"There are "<<phen_num<<" traits specified in the file ["+phen_file+"]."<<endl;
if(mphen>phen_num){
stringstream errmsg;
errmsg<<"Error: can not find the "<<mphen<<"th trait in the file ["+phen_file+"].";
throw(errmsg.str());
}
if(phen_num>1) cout<<"The "<<mphen<<"th trait is included for analysis."<<endl;
in_phen.seekg(ios::beg);
mphen--;
while(in_phen){
in_phen>>fid_buf;
if(in_phen.eof()) break;
in_phen>>pid_buf;
for(i=0; i<phen_num; i++){
in_phen>>str_buf;
if(i==mphen && str_buf!="-9" && str_buf!="NA"){
phen_ID.push_back(fid_buf+":"+pid_buf);
fid.push_back(fid_buf);
pid.push_back(pid_buf);
phen_buf.push_back(str_buf);
}
}
}
in_phen.close();
cout<<"Nonmissing phenotypes of "<<phen_buf.size()<<" individuals are included from ["+phen_file+"]."<<endl;
if(_id_map.empty()){
_fid=fid;
_pid=pid;
_indi_num=_fid.size();
init_keep();
}
}
void gcta::read_famfile(string famfile)
{
ifstream Fam(famfile.c_str());
if(!Fam) throw("Error: can not open the file ["+famfile+"] to read.");
cout<<"Reading PLINK FAM file from ["+famfile+"]."<<endl;
int i=0;
string str_buf;
_fid.clear();
_pid.clear();
_fa_id.clear();
_mo_id.clear();
_sex.clear();
_pheno.clear();
while(Fam){
Fam>>str_buf;
if(Fam.eof()) break;
_fid.push_back(str_buf);
Fam>>str_buf;
_pid.push_back(str_buf);
Fam>>str_buf;
_fa_id.push_back(str_buf);
Fam>>str_buf;
_mo_id.push_back(str_buf);
Fam>>str_buf;
_sex.push_back(atoi(str_buf.c_str()));
Fam>>str_buf;
_pheno.push_back(atoi(str_buf.c_str()));
}
Fam.clear();
Fam.close();
_indi_num=_fid.size();
cout<<_indi_num<<" individuals to be included from ["+famfile+"]."<<endl;
// Initialize _keep
init_keep();
}
void gcta::read_imp_dose_beagle(string zdosefile, string kp_indi_file, string rm_indi_file, string blup_indi_file)
{
if(_include.size()==0) throw("Error: No SNP is retained for analysis.");
int i=0, j=0;
vector<int> rsnp;
get_rsnp(rsnp);
const int MAX_LINE_LENGTH=10000000;
char buf[MAX_LINE_LENGTH];
string str_buf;
gzifstream zinf;
zinf.open( zdosefile.c_str() );
if(! zinf.is_open()) throw("Error: can not open the file ["+zdosefile+"] to read.");
cout<<"Reading imputed dosage scores (BEAGLE output) ..."<<endl;
zinf.getline(buf, MAX_LINE_LENGTH, '\n');
stringstream ss(buf);
for(i=0; i<3; i++) ss>>str_buf;
while(ss>>str_buf){ _fid.push_back(str_buf); }
_pid=_fid;
_indi_num=_fid.size();
_fa_id.resize(_indi_num);
_mo_id.resize(_indi_num);
_sex.resize(_indi_num);
_pheno.resize(_indi_num);
cout<<_indi_num<<" individuals to be included from ["+zdosefile+"]."<<endl;
init_keep();
if(!kp_indi_file.empty()) keep_indi(kp_indi_file);
if(!blup_indi_file.empty()) read_indi_blup(blup_indi_file);
if(!rm_indi_file.empty()) remove_indi(rm_indi_file);
_geno_dose.resize(_keep.size());
for(i=0; i<_keep.size(); i++) _geno_dose[i].resize(_include.size());
vector<int> rindi;
get_rindi(rindi);
int line=0;
int k=0;
double d_buf=0.0;
while(1){
zinf.getline(buf, MAX_LINE_LENGTH, '\n');
if(zinf.fail() || !zinf.good()) break;
if(!rsnp[line++]) continue;
stringstream ss(buf);
ss>>str_buf;
if(str_buf!=_snp_name[line-1]){
stringstream errmsg;
errmsg<<"Error: the "<<line<<" th SNP ["+_snp_name[line-1]+"] in the summary file doesn't match to that in the dosage file."<<endl;
throw(errmsg.str());
}
ss>>str_buf>>str_buf;
for(i=0, j=0; i<_indi_num; i++){
ss>>d_buf;
if(rindi[i]){ _geno_dose[j][k]=d_buf; j++; }
}
k++;
}
zinf.clear();
zinf.close();
}
void gcta::read_imp_dose_mach(string zdosefile, string kp_indi_file, string rm_indi_file, string blup_indi_file)
{
if(_include.size()==0) throw("Error: No SNP is retained for analysis.");
int i=0, j=0, k=0, line=0;
vector<int> rsnp;
get_rsnp(rsnp);
const int MAX_LINE_LENGTH = 10000000;
char buf[MAX_LINE_LENGTH];
gzifstream zinf;
zinf.open(zdosefile.c_str());
if(! zinf.is_open()) throw("Error: can not open the file ["+zdosefile+"] to read.");
vector<string> indi_ls;
map<string, int> kp_id_map, blup_id_map, rm_id_map;
bool kp_indi_flag=!kp_indi_file.empty(), blup_indi_flag=!blup_indi_file.empty(), rm_indi_flag=!rm_indi_file.empty();
if(kp_indi_flag) read_indi_list(kp_indi_file, indi_ls);
for(i=0; i<indi_ls.size(); i++) kp_id_map.insert(pair<string, int>(indi_ls[i], i));
if(blup_indi_flag) read_indi_list(blup_indi_file, indi_ls);
for(i=0; i<indi_ls.size(); i++) blup_id_map.insert(pair<string, int>(indi_ls[i], i));
if(rm_indi_flag) read_indi_list(rm_indi_file, indi_ls);
for(i=0; i<indi_ls.size(); i++) rm_id_map.insert(pair<string, int>(indi_ls[i], i));
bool missing=false;
string str_buf, id_buf, err_msg="Error: reading dosage data failed. Are the map file and the dosage file matched?";
double f_buf=0.0;
vector<string> kept_id, vs_buf;
cout<<"Reading dosage data from ["+zdosefile+"] in individual-major format (Note: may use huge RAM)."<<endl;
_fid.clear();
_pid.clear();
_geno_dose.clear();
vector<int> kp_it;
while(1){
bool kp_flag=true;
zinf.getline(buf, MAX_LINE_LENGTH, '\n');
stringstream ss(buf);
if(!(ss>>str_buf)) break;
int ibuf=StrFunc::split_string(str_buf, vs_buf, ">");
if(ibuf>1){
if(vs_buf[0].empty()) throw("Error: family ID of the individual ["+str_buf+"] is missing.");
else vs_buf[0].erase(vs_buf[0].end()-1);
}
else if(ibuf==1) vs_buf.push_back(vs_buf[0]);
else break;
id_buf=vs_buf[0]+":"+vs_buf[1];
if(kp_indi_flag && kp_id_map.find(id_buf)==kp_id_map.end()) kp_flag=false;
if(kp_flag && blup_indi_flag && blup_id_map.find(id_buf)==blup_id_map.end()) kp_flag=false;
if(kp_flag && rm_indi_flag && rm_id_map.find(id_buf)!=rm_id_map.end()) kp_flag=false;
if(kp_flag){
kp_it.push_back(1);
_fid.push_back(vs_buf[0]);
_pid.push_back(vs_buf[1]);
kept_id.push_back(id_buf);
}
else kp_it.push_back(0);
if(zinf.fail() || !zinf.good()) break;
}
zinf.clear();
zinf.close();
cout<<"(Imputed dosage data for "<<kp_it.size()<<" individuals detected)."<<endl;
_indi_num=_fid.size();
zinf.open(zdosefile.c_str());
_geno_dose.resize(_indi_num);
for(line=0; line<_indi_num; line++) _geno_dose[line].resize(_include.size());
for(line=0, k=0; line<kp_it.size(); line++){
zinf.getline(buf, MAX_LINE_LENGTH, '\n');
if(kp_it[line]==0) continue;
stringstream ss(buf);
if(!(ss>>str_buf)) break;
if(!(ss>>str_buf)) break;
for(i=0, j=0; i<_snp_num; i++){
ss>>str_buf;
f_buf=atof(str_buf.c_str());
if(str_buf=="X" || str_buf=="NA"){
if(!missing){
cout<<"Warning: missing values detected in the dosage data."<<endl;
missing=true;
}
f_buf=1e6;
}
if(rsnp[i]){ _geno_dose[k][j]=(f_buf); j++; }
}
k++;
}
zinf.clear();
zinf.close();
cout<<"Imputed dosage data for "<<kept_id.size()<<" individuals are included from ["<<zdosefile<<"]."<<endl;
_fa_id.resize(_indi_num);
_mo_id.resize(_indi_num);
_sex.resize(_indi_num);
_pheno.resize(_indi_num);
for(i=0; i<_indi_num; i++){
_fa_id[i]=_mo_id[i]="0";
_sex[i]=-9;
_pheno[i]=-9;
}
// initialize keep
init_keep();
update_id_map_kp(kept_id, _id_map, _keep);
if(_keep.size()==0) throw("Error: No individual is retained for analysis.");
if(blup_indi_flag) read_indi_blup(blup_indi_file);
// update data
update_bim(rsnp);
}
