// *** MAIN START***
int main (int argc, const char * argv[]){
cmdl_opts opts;
get_opts( argc, argv, opts);
int nTimes=0, nT=0;
//*** EMITTED DATA OBJECTS ***
Emission cnaEmit, bafEmit, snvEmit;
if (opts.cna_fn != NULL) get_cna_data( &cnaEmit, opts, nTimes);
if (opts.baf_fn != NULL) get_baf_data( &bafEmit, opts, nTimes, nT);
if (opts.snv_fn != NULL) get_snv_data( &snvEmit, opts, nTimes, nT);
//*** ANNOUNCE ***
printf("\ncloneHD: probabilistic inference of sub-clonality using...\n\n");
if (cnaEmit.is_set){
printf("CNA data in %s: %i sites in %i chr across %i samples\n",
opts.cna_fn, cnaEmit.total_loci, cnaEmit.nSamples, nTimes);
}
if (bafEmit.is_set){
printf("BAF data in %s: %i sites in %i chr across %i samples\n",
opts.baf_fn, bafEmit.total_loci, bafEmit.nSamples, nTimes);
}
if (snvEmit.is_set){
printf("SNV data in %s: %i sites in %i chr across %i samples\n",
opts.snv_fn, snvEmit.total_loci, snvEmit.nSamples, nTimes);
}
cout<<endl;
// *** ALLOCATE CLONE ***
Clone myClone;
myClone.allocate( &cnaEmit, &bafEmit, &snvEmit, opts.chr_fn);
myClone.cna_pen_zero = opts.cna_pen_zero;//CNA penalty for zero total copies
myClone.cna_pen_diff = opts.cna_pen_diff;//CNA penalty for different c.n.
myClone.cna_pen_norm = opts.cna_pen_norm;//CNA penalty for non-normal c.n.
myClone.baf_pen_comp = opts.baf_pen_comp;//BAF penalty for complex chr status
myClone.snv_pen_high = opts.snv_pen_high;//SNV penalty for high SNV genotypes
myClone.snv_pen_mult = opts.snv_pen_mult;//SNV penalty for multiple hit SNVs
myClone.snv_fpr = opts.snv_fpr;//SNV false-positive rate
myClone.snv_fpf = opts.snv_fpf;//SNV frequency of false positives
myClone.bulk_fix = opts.bulk_fix;
myClone.cnaGrid = opts.cnaGrid;
myClone.bafGrid = opts.bafGrid;
myClone.snvGrid = opts.snvGrid;
myClone.bulkGrid = opts.bulkGrid;
myClone.learn_priors = (cnaEmit.is_set || snvEmit.connect || opts.avcn_fn != NULL) ? 0 : opts.learn_priors;
// *** GET MAX-TCN INFO ***
get_maxtcn_input( opts.maxtcn_fn, opts.maxtcn, &myClone);
// *** GET SNV BULK PRIOR ***
if ( snvEmit.is_set && opts.bulk_fn != NULL ){
printf("Using data in %s as SNV bulk prior...\n", opts.bulk_fn);
get_snv_bulk_prior( &myClone, opts);
}
//*** GET JUMP PROBABILITY TRACKS and COLLAPSE TO EVENTS***
get_jump_probability( &myClone, opts);
//...now all segments are fixed and mean_tcn/av_cn allocated.
if ( snvEmit.is_set && !cnaEmit.is_set ){//for SNV only
// *** GET TOTAL MEAN COPYNUMBER TRACK ***
if( opts.mntcn_fn != NULL ){
get_mean_tcn( opts.mntcn_fn, &myClone, &snvEmit);
}
// *** GET AVAILABLE COPYNUMBER TRACK ***
if ( opts.avcn_fn != NULL ){
get_avail_cn( opts.avcn_fn, &myClone, &snvEmit);
}
}
//*** GET READ DEPTH BIAS FIELD ***
if (cnaEmit.is_set && opts.bias_fn != NULL){
get_bias_field( &myClone, opts);
}
//*** PREPARE COARSE-GRAINED DATA ***
if (cnaEmit.is_set && (opts.cna_jumps_fn != NULL || opts.cna_jump == 0.0)){
cnaEmit.log_space = 1;
cnaEmit.coarse_grained = 1;
printf( "Collapsed CNA data to %5i segments based on potential jump events.\n",
cnaEmit.total_events);
cout<<"Precomputing for CNA..."<<flush;
myClone.get_cnaEmitLog();
cout<<"done."<<endl;
}
if (bafEmit.is_set && ( opts.cna_jumps_fn != NULL || opts.baf_jumps_fn != NULL || opts.baf_jump == 0.0)){
bafEmit.log_space = 1;
bafEmit.coarse_grained = 1;
printf("Collapsed BAF data to %5i segments based on potential jump events.\n", bafEmit.total_events);
cout<<"Precomputing for BAF..."<<flush;
myClone.get_bafEmitLog();
cout<<"done."<<endl;
}
if (snvEmit.is_set && opts.snv_jumps_fn != NULL){
snvEmit.log_space = 1;
snvEmit.coarse_grained = 1;
printf("Collapsed SNV data to %5i segments based on potential jump events.\n", snvEmit.total_events);
cout<<"Precomputing for SNV..."<<flush;
myClone.get_snvEmitLog();
cout<<"done."<<endl;
}
cout<<endl;
//exit(0);
// get purities...
if (opts.purity_fn != NULL){
get_purity( opts.purity_fn, myClone.min_purity);
}
// get user pre-defined clones
gsl_matrix * clones = NULL;
gsl_vector * mass = NULL;
if (opts.clones_fn != NULL) get_fixed_clones( clones, mass, opts.clones_fn, nTimes);
int bestn=0, rows=0;
if (mass != NULL && (int) mass->size > nTimes) rows = (int) mass->size;
if (clones != NULL && (int) clones->size1 > nTimes) rows = (int) clones->size1;
if (rows > nTimes){//print LLH's for predefined parameter values...
print_llh_for_set( clones, mass, &myClone, opts);
return(0);
}
else{
// ****** INFERENCE STARTS HERE ******
bestn = infer_clones( clones, mass, &myClone, opts);
printf("cloneHD in ");
if (cnaEmit.is_set && bafEmit.is_set && snvEmit.is_set) cout<<"cna-baf-snv ";
if (cnaEmit.is_set && bafEmit.is_set && !snvEmit.is_set) cout<<"cna-baf ";
if (cnaEmit.is_set && !bafEmit.is_set && snvEmit.is_set) cout<<"cna-snv ";
if (cnaEmit.is_set && !bafEmit.is_set && !snvEmit.is_set) cout<<"cna ";
if (!cnaEmit.is_set && !bafEmit.is_set && snvEmit.is_set) cout<<"snv ";
printf("mode found support for %i sub-clone(s) in the data.\n", bestn);
// ****** INFERENCE COMPLETED ********
}
print_all_results( &myClone, opts);
// all done...
return (0);
}
