//chr start stop is given from commandine
myMap::iterator iter_init(perpsmc *pp,char *chr,int start,int stop){
assert(chr!=NULL);
myMap::iterator it = pp->mm.find(chr);
if(it==pp->mm.end()){
fprintf(stderr,"\t-> [%s] Problem finding chr: %s\n",__FUNCTION__,chr);
return it;
}
my_bgzf_seek(pp->bgzf_gls,it->second.saf,SEEK_SET);
my_bgzf_seek(pp->bgzf_pos,it->second.pos,SEEK_SET);
//fprintf(stderr,"pp->gls:%p\n",pp->gls);
if(pp->pos)
delete [] pp->pos;
if(pp->gls)
delete [] pp->gls;
pp->pos = new int[it->second.nSites];
my_bgzf_read(pp->bgzf_pos,pp->pos,sizeof(int)*it->second.nSites);
pp->gls = new double[2*it->second.nSites];//<-valgrind complains about large somthing
my_bgzf_read(pp->bgzf_gls,pp->gls,2*sizeof(double)*it->second.nSites);
// fprintf(stderr," end: %f %f\n",pp->gls[0],pp->gls[1]);
pp->first=0;
if(start!=-1)
while(pp->first<it->second.nSites&&pp->pos[pp->first]<start)
pp->first++;
pp->last = it->second.nSites;
if(stop!=-1&&stop<=pp->pos[pp->last-1]){
pp->last=pp->first;
while(pp->pos[pp->last]<stop)
pp->last++;
}
return it;
}
perpsmc * perpsmc_init(char *fname){
perpsmc *ret = new perpsmc ;
ret->fname = strdup(fname);
ret->gls =NULL;
ret->pos = NULL;
ret->bgzf_pos=ret->bgzf_gls=NULL;
ret->pos = NULL;
size_t clen;
if(!fexists(fname)){
fprintf(stderr,"\t-> Problem opening file: \'%s\'\n",fname);
exit(0);
}
FILE *fp = NULL;
fp=fopen(fname,"r");
if(fp==NULL){
fprintf(stderr,"\t-> Problem opening file:%s\n",fname);
exit(0);
}
char buf[8];
assert(fread(buf,1,8,fp)==8);
ret->version = psmcversion(fname);
fprintf(stderr,"\t-> Version of fname: \'%s\' is:%d\n",fname,ret->version);
if(ret->version!=1){
fprintf(stderr,"\t-> Looks like you are trying to use a version of PSMC that does not exists\n");
exit(0);
}
ret->nSites =0;
while(fread(&clen,sizeof(size_t),1,fp)){
char *chr = (char*)malloc(clen+1);
assert(clen==fread(chr,1,clen,fp));
chr[clen] = '\0';
datum d;
if(1!=fread(&d.nSites,sizeof(size_t),1,fp)){
fprintf(stderr,"[%s.%s():%d] Problem reading data: %s \n",__FILE__,__FUNCTION__,__LINE__,fname);
exit(0);
}
ret->nSites += d.nSites;
if(1!=fread(&d.pos,sizeof(int64_t),1,fp)){
fprintf(stderr,"[%s->%s():%d] Problem reading data: %s \n",__FILE__,__FUNCTION__,__LINE__,fname);
exit(0);
}
if(1!=fread(&d.saf,sizeof(int64_t),1,fp)){
fprintf(stderr,"[%s->%s():%d] Problem reading data: %s \n",__FILE__,__FUNCTION__,__LINE__,fname);
exit(0);
}
myMap::iterator it = ret->mm.find(chr);
if(it==ret->mm.end())
ret->mm[chr] =d ;
else{
fprintf(stderr,"Problem with chr: %s, key already exists, psmc file needs to be sorted. (sort your -rf that you used for input)\n",chr);
exit(0);
}
}
fclose(fp);
char *tmp =(char*)calloc(strlen(fname)+100,1);//that should do it
tmp=strncpy(tmp,fname,strlen(fname)-3);
// fprintf(stderr,"tmp:%s\n",tmp);
char *tmp2 = (char*)calloc(strlen(fname)+100,1);//that should do it
snprintf(tmp2,strlen(fname)+100,"%sgz",tmp);
fprintf(stderr,"\t-> Assuming .psmc.gz file: %s\n",tmp2);
ret->bgzf_gls = bgzf_open(tmp2,"r");
if(ret->bgzf_gls)
my_bgzf_seek(ret->bgzf_gls,8,SEEK_SET);
if(ret->bgzf_gls && ret->version!=psmcversion(tmp2)){
fprintf(stderr,"\t-> Problem with mismatch of version of %s vs %s %d vs %d\n",fname,tmp2,ret->version,psmcversion(tmp2));
exit(0);
}
snprintf(tmp2,strlen(fname)+100,"%spos.gz",tmp);
fprintf(stderr,"\t-> Assuming .psmc.pos.gz: %s\n",tmp2);
ret->bgzf_pos = bgzf_open(tmp2,"r");
if(ret->pos)
my_bgzf_seek(ret->bgzf_pos,8,SEEK_SET);
if(ret->bgzf_pos&& ret->version!=psmcversion(tmp2)){
fprintf(stderr,"Problem with mismatch of version of %s vs %s\n",fname,tmp2);
exit(0);
}
//assert(ret->pos!=NULL&&ret->saf!=NULL);
free(tmp);free(tmp2);
return ret;
}
int fst_stat(int argc,char **argv){
char *bname = *argv;
fprintf(stderr,"\t-> Assuming idxname:%s\n",bname);
perfst *pf = perfst_init(bname);
args *pars = getArgs(--argc,++argv);
int *ppos = NULL;
int chs = choose((int)pf->names.size(),2);
// fprintf(stderr,"choose:%d \n",chs);
double **ares = new double*[chs];
double **bres = new double*[chs];
double unweight[chs];
double wa[chs];
double wb[chs];
size_t nObs[chs];
for(int i=0;i<chs;i++){
unweight[i] = wa[i] = wb[i] =0.0;
nObs[i] = 0;
}
for(myFstMap::iterator it=pf->mm.begin();it!=pf->mm.end();++it){
if(pars->chooseChr!=NULL){
it = pf->mm.find(pars->chooseChr);
if(it==pf->mm.end()){
fprintf(stderr,"Problem finding chr: %s\n",pars->chooseChr);
break;
}
}
if(it->second.nSites==0)
continue;
my_bgzf_seek(pf->fp,it->second.off,SEEK_SET);
ppos = new int[it->second.nSites];
my_bgzf_read(pf->fp,ppos,sizeof(int)*it->second.nSites);
for(int i=0;i<choose((int)pf->names.size(),2);i++){
ares[i] = new double[it->second.nSites];
bres[i] = new double[it->second.nSites];
my_bgzf_read(pf->fp,ares[i],sizeof(double)*it->second.nSites);
my_bgzf_read(pf->fp,bres[i],sizeof(double)*it->second.nSites);
}
size_t first=0;
if(pars->start!=-1)
while(ppos[first]<pars->start)
first++;
size_t last=it->second.nSites;
if(pars->stop!=-1&&pars->stop<=ppos[last-1]){
last=first;
while(ppos[last]<pars->stop)
last++;
}
// fprintf(stderr,"pars->stop:%d ppos:%d first:%d last:%d\n",pars->stop,ppos[last-1],first,last);
for(size_t s=first;s<last;s++){
#if 0
fprintf(stdout,"%s\t%d",it->first,ppos[s]+1);
for(int i=0;i<choose(pf->names.size(),2);i++)
fprintf(stdout,"\t%f\t%f",ares[i][s],bres[i][s]);
fprintf(stdout,"\n");
#endif
for(int i=0;i<choose((int)pf->names.size(),2);i++){
if(bres[i][s]!=0){
unweight[i] += ares[i][s]/bres[i][s];
nObs[i]++;
}
wa[i] += ares[i][s];
wb[i] += bres[i][s];
}
}
for(int i=0;i<choose((int)pf->names.size(),2);i++){
delete [] ares[i];
delete [] bres[i];
}
delete [] ppos;
if(pars->chooseChr!=NULL)
break;
}
double fstUW[chs];
double fstW[chs];
for(int i=0;i<chs;i++){
fstUW[i] = unweight[i]/(1.0*((double)nObs[i]));
fstW[i] = wa[i]/wb[i];
fprintf(stderr,"\t-> FST.Unweight[nObs:%lu]:%f Fst.Weight:%f\n",nObs[i],fstUW[i],fstW[i]);
fprintf(stdout,"%f %f\n",fstUW[i],fstW[i]);
}
if(chs==3){
//if chr==3 then we have 3pops and we will also calculate pbs statistics
calcpbs(fstW);//<- NOTE: the pbs values will replace the fstW values
for(int i=0;i<3;i++)
fprintf(stderr,"\t-> pbs.pop%d\t%f\n",i+1,fstW[i]);
}
delete [] ares;
delete [] bres;
destroy_args(pars);
perfst_destroy(pf);
return 0;
}
int fst_print(int argc,char **argv){
char *bname = *argv;
fprintf(stderr,"\t-> Assuming idxname:%s\n",bname);
perfst *pf = perfst_init(bname);
writefst_header(stderr,pf);
args *pars = getArgs(--argc,++argv);
int *ppos = NULL;
fprintf(stderr,"choose:%d \n",choose((int)pf->names.size(),2));
double **ares = new double*[choose((int)pf->names.size(),2)];
double **bres = new double*[choose((int)pf->names.size(),2)];
for(myFstMap::iterator it=pf->mm.begin();it!=pf->mm.end();++it){
if(pars->chooseChr!=NULL){
it = pf->mm.find(pars->chooseChr);
if(it==pf->mm.end()){
fprintf(stderr,"Problem finding chr: %s\n",pars->chooseChr);
break;
}
}
if(it->second.nSites==0)
continue;
my_bgzf_seek(pf->fp,it->second.off,SEEK_SET);
ppos = new int[it->second.nSites];
my_bgzf_read(pf->fp,ppos,sizeof(int)*it->second.nSites);
for(int i=0;i<choose((int)pf->names.size(),2);i++){
ares[i] = new double[it->second.nSites];
bres[i] = new double[it->second.nSites];
my_bgzf_read(pf->fp,ares[i],sizeof(double)*it->second.nSites);
my_bgzf_read(pf->fp,bres[i],sizeof(double)*it->second.nSites);
}
size_t first=0;
if(pars->start!=-1)
while(ppos[first]<pars->start)
first++;
size_t last=it->second.nSites;
if(pars->stop!=-1&&pars->stop<=ppos[last-1]){
last=first;
while(ppos[last]<pars->stop)
last++;
}
fprintf(stderr,"pars->stop:%d ppos:%d first:%lu last:%lu\n",pars->stop,ppos[last-1],first,last);
for(size_t s=first;s<last;s++){
fprintf(stdout,"%s\t%d",it->first,ppos[s]+1);
for(int i=0;i<choose((int)pf->names.size(),2);i++)
fprintf(stdout,"\t%f\t%f",ares[i][s],bres[i][s]);
fprintf(stdout,"\n");
}
for(int i=0;i<choose((int)pf->names.size(),2);i++){
delete [] ares[i];
delete [] bres[i];
}
delete [] ppos;
if(pars->chooseChr!=NULL)
break;
}
delete [] ares;
delete [] bres;
destroy_args(pars);
perfst_destroy(pf);
return 0;
}
int fst_stat2(int argc,char **argv){
int pS,pE;//physical start,physical end
int begI,endI;//position in array for pS, pE;
char *bname = *argv;
fprintf(stderr,"\t-> Assuming idxname:%s\n",bname);
perfst *pf = perfst_init(bname);
args *pars = getArgs(--argc,++argv);
fprintf(stderr,"win:%d step:%d\n",pars->win,pars->step);
int *ppos = NULL;
int chs = choose((int)pf->names.size(),2);
// fprintf(stderr,"choose:%d \n",chs);
double **ares = new double*[chs];
double **bres = new double*[chs];
double unweight[chs];
double wa[chs];
double wb[chs];
size_t nObs =0;
//print header
fprintf(stdout,"region\tchr\tmidPos\tNsites");
for(int c1=0;c1<chs-1;c1++)
for(int c2=c1+1;c2<chs;c2++)
fprintf(stdout,"\tFst%d%d",c1,c2);
if(chs==3)
fprintf(stdout,"\tPBS0\tPBS1\tPBS2");
fprintf(stdout,"\n");
for(myFstMap::iterator it=pf->mm.begin();it!=pf->mm.end();++it){
if(pars->chooseChr!=NULL){
it = pf->mm.find(pars->chooseChr);
if(it==pf->mm.end()){
fprintf(stderr,"Problem finding chr: %s\n",pars->chooseChr);
break;
}
}
fprintf(stderr,"nSites:%lu\n",it->second.nSites);
if(it->second.nSites==0&&pars->chooseChr!=NULL)
break;
else if(it->second.nSites==0&&pars->chooseChr==NULL)
continue;
my_bgzf_seek(pf->fp,it->second.off,SEEK_SET);
ppos = new int[it->second.nSites];
my_bgzf_read(pf->fp,ppos,sizeof(int)*it->second.nSites);
for(int i=0;i<it->second.nSites;i++)//what? why? dragon!
ppos[i]++;
for(int i=0;i<choose((int)pf->names.size(),2);i++){
ares[i] = new double[it->second.nSites];
bres[i] = new double[it->second.nSites];
my_bgzf_read(pf->fp,ares[i],sizeof(double)*it->second.nSites);
my_bgzf_read(pf->fp,bres[i],sizeof(double)*it->second.nSites);
}
if(pars->type==0)
pS = ((pars->start!=-1?pars->start:ppos[0])/pars->step)*pars->step +pars->step;
else if(pars->type==1)
pS = (pars->start!=-1?pars->start:ppos[0]);
else if(pars->type==2)
pS = 1;
pE = pS+pars->win;
begI=endI=0;
//fprintf(stderr,"ps:%d\n",pS);exit(0);
if(pE>(pars->stop!=-1?pars->stop:ppos[it->second.nSites-1])){
fprintf(stderr,"end of dataset is before end of window: end of window:%d last position in chr:%d\n",pE,ppos[it->second.nSites-1]);
// return str;
}
while(ppos[begI]<pS) begI++;
endI=begI;
while(endI<it->second.nSites &&ppos[endI]<pE) endI++;
//fprintf(stderr,"begI:%d endI:%d\n",begI,endI);
while(1){
for(int i=0;i<chs;i++)
unweight[i] = wa[i] = wb[i] =0.0;
nObs=0;
for(int s=begI;s<endI;s++){
#if 0
fprintf(stdout,"%s\t%d",it->first,ppos[s]+1);
for(int i=0;i<choose(pf->names.size(),2);i++)
fprintf(stdout,"\t%f\t%f",ares[i][s],bres[i][s]);
fprintf(stdout,"\n");
#endif
for(int i=0;i<choose((int)pf->names.size(),2);i++){
unweight[i] += ares[i][s]/bres[i][s];
wa[i] += ares[i][s];
wb[i] += bres[i][s];
}
nObs++;
}
if(nObs>0)
fprintf(stdout,"(%d,%d)(%d,%d)(%d,%d)\t%s\t%d\t%d",begI,endI-1,ppos[begI],ppos[endI-1],pS,pE,it->first,pS+(pE-pS)/2,endI-begI+1);
double fstW[chs];
for(int i=0;nObs>0&&i<chs;i++){
fstW[i] = wa[i]/wb[i];
// fprintf(stdout,"\t%f\t%f",unweight[i]/(1.0*nObs),fstW[i]);
fprintf(stdout,"\t%f",fstW[i]);
}
if(nObs>0&&chs==3){
//if chr==3 then we have 3pops and we will also calculate pbs statistics
calcpbs(fstW);//<- NOTE: the pbs values will replace the fstW values
for(int i=0;i<3;i++)
fprintf(stdout,"\t%f",fstW[i]);
}
if(nObs>0)
fprintf(stdout,"\n");
pS += pars->step;
pE =pS+pars->win;
if(pE>(pars->stop!=-1?pars->stop:ppos[it->second.nSites-1]))
break;
while(ppos[begI]<pS) begI++;
while(ppos[endI]<pE) endI++;
}
for(int i=0;i<choose((int)pf->names.size(),2);i++){
delete [] ares[i];
delete [] bres[i];
}
delete [] ppos;
if(pars->chooseChr!=NULL)
break;
}
delete [] ares;
delete [] bres;
destroy_args(pars);
perfst_destroy(pf);
return 0;
}