int oneBin
(int *profile_chromStart,
int *profile_chromEnd,
int *profile_coverage,
int n_profiles,
int *bin_total,
int bin_chromStart,
int bin_chromEnd){
return binSum(
profile_chromStart,
profile_chromEnd,
profile_coverage,
n_profiles,
bin_total,
bin_chromEnd - bin_chromStart,
1,
bin_chromStart,
EMPTY_AS_ZERO);
}
int
binSumLR
(int *data_start_end,
int *chromStart, int *chromEnd,
int *coverage, int n_entries,
int *left_bin_vec, int *right_bin_vec,
int left_chromStart, int right_chromStart,
int bases_per_bin, int n_bins){
int bin_chromEnd, bin_chromStart;
int extra_chromStart, extra_chromEnd, extra_bases, extra_coverage;
int status;
/* printf("left bin_size=%d bins=%d start=%d\n", */
/* bases_per_bin, n_bins, left_chromStart); */
status = binSum(chromStart, chromEnd,
coverage, n_entries,
left_bin_vec,
bases_per_bin,
n_bins,
left_chromStart,
EMPTY_AS_ZERO);
if(status != 0){
return status;
}
/* printf("right bin_size=%d bins=%d start=%d\n", */
/* bases_per_bin, n_bins, right_chromStart); */
status = binSum(chromStart, chromEnd,
coverage, n_entries,
right_bin_vec,
bases_per_bin,
n_bins,
right_chromStart,
EMPTY_AS_ZERO);
if(status != 0){
return status;
}
for(int bin_i=0; bin_i < n_bins; bin_i++){
//left bin.
bin_chromStart = left_chromStart + bases_per_bin * bin_i;
bin_chromEnd = bin_chromStart + bases_per_bin;
if(data_start_end[0] < bin_chromEnd){
if(data_start_end[0] <= bin_chromStart){
// ( data ]
// (bin]
// (bin]
// bin is completely data, leave it alone!
}else{
// ( data ]
// (bin]
// - extra
// (bin]
// --- extra
// bin has some data, so subtract the extra.
extra_chromStart = bin_chromStart;
extra_chromEnd = data_start_end[0];
extra_bases = extra_chromEnd - extra_chromStart;
//printf("left start=%d bases=%d\n", extra_chromStart, extra_bases);
status = binSum(chromStart, chromEnd,
coverage, n_entries,
&extra_coverage,
extra_bases,
1,
extra_chromStart,
EMPTY_AS_ZERO);
if(status != 0){
return status;
}
left_bin_vec[bin_i] -= extra_coverage;
}
}else{
// ( data ]
// (bin]
// bin does not overlap data, so set it to zero.
left_bin_vec[bin_i] = 0;
}
//right bin.
bin_chromStart = right_chromStart + bases_per_bin * bin_i;
bin_chromEnd = bin_chromStart + bases_per_bin;
if(bin_chromStart < data_start_end[1]){
if(bin_chromEnd <= data_start_end[1]){
// ( data ]
// (bin]
// (bin]
// bin is completely data, leave it alone!
}else{
// ( data ]
// (bin]
// -- extra
extra_chromStart = data_start_end[1];
extra_chromEnd = bin_chromEnd;
extra_bases = extra_chromEnd - extra_chromStart;
//printf("right start=%d bases=%d\n", extra_chromStart, extra_bases);
status = binSum(chromStart, chromEnd,
coverage, n_entries,
&extra_coverage,
extra_bases,
1,
extra_chromStart,
EMPTY_AS_ZERO);
if(status != 0){
return status;
}
right_bin_vec[bin_i] -= extra_coverage;
}
}else{
// ( data ]
// (bin]
// (bin]
// bin does not overlap data, so set it to zero.
right_bin_vec[bin_i] = 0;
}
}
return 0;
}
int PeakSegJointFaster(
struct ProfileList *profile_list,
int bin_factor,
double *mean_mat,
double *flat_loss_vec,
double *peak_loss_vec,
int *peak_start_end,
int *data_start_end
){
int n_samples = profile_list->n_profiles;
if(n_samples == 0){
return ERROR_FASTER_NO_COVERAGE_DATA;
}
int chromStart, chromEnd, unfilled_chromStart, unfilled_chromEnd;
double seg1_mean=-1, seg2_mean, seg3_mean;
struct Profile *profile, *samples = profile_list->profile_vec;
profile = samples;
unfilled_chromEnd = get_max_chromEnd(profile);
unfilled_chromStart = get_min_chromStart(profile);
for(int sample_i=1; sample_i < n_samples; sample_i++){
profile = samples + sample_i;
chromStart = get_min_chromStart(profile);
if(chromStart < unfilled_chromStart){
unfilled_chromStart = chromStart;
}
chromEnd = get_max_chromEnd(profile);
if(unfilled_chromEnd < chromEnd){
unfilled_chromEnd = chromEnd;
}
}
data_start_end[0] = unfilled_chromStart;
data_start_end[1] = unfilled_chromEnd;
int unfilled_bases = unfilled_chromEnd - unfilled_chromStart;
double data_bases = (double) unfilled_bases;
double bin_bases;
if(unfilled_bases/bin_factor < 4){
/*
4 is smallest the number of data points for which the 3-segment
optimization problem is not trivial.
If we don't have at least this many data points for the first
bin step, than we stop with an error.
*/
return ERROR_FASTER_BIN_FACTOR_TOO_LARGE;
}
int bases_per_bin = 1;
while(unfilled_bases/bases_per_bin/bin_factor >= 4){
bases_per_bin *= bin_factor;
}
int n_bins = unfilled_bases / bases_per_bin;
/*
MaxBinSize() from line 1 of the JointZoom algorithm of the
PeakSegJoint paper returns the value of the C variable
bases_per_bin.
Little b in the text of the section that describes the JointZoom
algorithm is the C variable n_bins.
*/
if(unfilled_bases % bases_per_bin != 0){
n_bins ++ ;
}
if(n_bins < bin_factor*2){
n_bins = bin_factor*2;
}
int extra_bases = n_bins * bases_per_bin - unfilled_bases;
int extra_before = extra_bases/2;
int extra_after = extra_bases - extra_before;
//Rprintf("bin_factor=%d n_bins=%d bases_per_bin=%d extra_before=%d extra_after=%d\n", bin_factor, n_bins, bases_per_bin, extra_before, extra_after);
//int extra_count;
int seg1_chromStart = unfilled_chromStart - extra_before;
int seg3_chromEnd = unfilled_chromEnd + extra_after;
int *sample_count_mat = (int*) malloc(n_bins * n_samples * sizeof(int));
double *last_cumsum_vec = (double*) malloc(n_samples*sizeof(double));
int *sample_cumsum_mat = (int*) malloc(n_bins * n_samples * sizeof(int));
double *seg1_mean_vec = (double*)malloc(sizeof(double)*n_samples);
double *seg2_mean_vec = (double*)malloc(sizeof(double)*n_samples);
double *seg3_mean_vec = (double*)malloc(sizeof(double)*n_samples);
double *seg1_loss_vec = (double*)malloc(sizeof(double)*n_samples);
double *candidate_loss_vec = (double*)malloc(sizeof(double)*n_samples);
int *left_bin_vec = (int*) malloc(n_bins * sizeof(int));
int *right_bin_vec = (int*) malloc(n_bins * sizeof(int));
int *left_cumsum_mat = (int*) malloc(n_bins * n_samples * sizeof(int));
int *right_cumsum_mat = (int*) malloc(n_bins * n_samples * sizeof(int));
int *left_initial_cumsum_vec = (int*) malloc(n_samples * sizeof(int));
int *right_initial_cumsum_vec = (int*) malloc(n_samples * sizeof(int));
int *left_cumsum_vec, *right_cumsum_vec;
int left_cumsum_value, right_cumsum_value;
int peakStart, peakEnd;
int best_seg1, best_seg2;
int *count_vec, *cumsum_vec, cumsum_value;
int status;
for(int sample_i=0; sample_i < n_samples; sample_i++){
profile = samples + sample_i;
count_vec = sample_count_mat + n_bins*sample_i;
status = binSum(profile->chromStart, profile->chromEnd,
profile->coverage, profile->n_entries,
count_vec,
bases_per_bin, n_bins, seg1_chromStart,
EMPTY_AS_ZERO);
if(status != 0){
free(sample_count_mat);
free(last_cumsum_vec);
free(sample_cumsum_mat);
free(seg1_mean_vec);
free(seg2_mean_vec);
free(seg3_mean_vec);
free(seg1_loss_vec);
free(candidate_loss_vec);
free(left_bin_vec);
free(right_bin_vec);
free(left_cumsum_mat);
free(right_cumsum_mat);
free(left_initial_cumsum_vec);
free(right_initial_cumsum_vec);
return status;
}
}//for sample_i
int bin_i, offset;
for(int sample_i=0; sample_i < n_samples; sample_i++){
cumsum_value = 0;
offset = n_bins * sample_i;
count_vec = sample_count_mat + offset;
cumsum_vec = sample_cumsum_mat + offset;
//printf("[sample%02d] ", sample_i);
for(bin_i=0; bin_i < n_bins; bin_i++){
cumsum_value += count_vec[bin_i];
//printf("%d ", cumsum_value);
cumsum_vec[bin_i] = cumsum_value;
}
last_cumsum_vec[sample_i] = cumsum_value;
seg1_mean = cumsum_value / data_bases;
flat_loss_vec[sample_i] = OptimalPoissonLoss(cumsum_value, seg1_mean);
}
/*
The for loops below implement the GridSearch() function mentioned
on line 2 of the JointZoom algorithm in the PeakSegJoint paper.
*/
double best_loss = INFINITY, feasible_loss;
for(int seg1_LastIndex=0; seg1_LastIndex < n_bins-2; seg1_LastIndex++){
peakStart = seg1_chromStart + (seg1_LastIndex+1)*bases_per_bin;
if(unfilled_chromStart < peakStart){
for(int sample_i=0; sample_i < n_samples; sample_i++){
cumsum_vec = sample_cumsum_mat + n_bins*sample_i;
cumsum_value = cumsum_vec[seg1_LastIndex];
bin_bases = (seg1_LastIndex+1)*bases_per_bin;
data_bases = bin_bases - (double)extra_before;
/* printf("sample_i=%d extra_before=%d bin_bases=%f data_bases=%f\n", */
/* sample_i, extra_before, bin_bases, data_bases); */
seg1_mean = cumsum_value/data_bases;
seg1_mean_vec[sample_i] = seg1_mean;
seg1_loss_vec[sample_i] = OptimalPoissonLoss(cumsum_value, seg1_mean);
}
for(int seg2_LastIndex=seg1_LastIndex+1;
seg2_LastIndex < n_bins-1;
seg2_LastIndex++){
feasible_loss=0.0;
peakEnd = seg1_chromStart + (seg2_LastIndex+1)*bases_per_bin;
if(peakEnd < unfilled_chromEnd){
for(int sample_i=0; sample_i < n_samples; sample_i++){
candidate_loss_vec[sample_i] = seg1_loss_vec[sample_i];
cumsum_vec = sample_cumsum_mat + n_bins*sample_i;
//segment 2.
cumsum_value = cumsum_vec[seg2_LastIndex]-cumsum_vec[seg1_LastIndex];
data_bases = (seg2_LastIndex-seg1_LastIndex)*bases_per_bin;
seg2_mean = cumsum_value/data_bases;
seg2_mean_vec[sample_i] = seg2_mean;
candidate_loss_vec[sample_i] += OptimalPoissonLoss(cumsum_value, seg2_mean);
//segment 3.
cumsum_value = cumsum_vec[n_bins-1]-cumsum_vec[seg2_LastIndex];
bin_bases = (n_bins-1-seg2_LastIndex)*bases_per_bin;
data_bases = bin_bases - extra_after;
seg3_mean = cumsum_value/data_bases;
seg3_mean_vec[sample_i] = seg3_mean;
candidate_loss_vec[sample_i] += OptimalPoissonLoss(cumsum_value, seg3_mean);
if(seg1_mean < seg2_mean && seg2_mean > seg3_mean){
feasible_loss += candidate_loss_vec[sample_i];
}else{
feasible_loss += flat_loss_vec[sample_i];
}
}//sample_i
if(feasible_loss < best_loss){
best_loss = feasible_loss;
peak_start_end[0] = peakStart;
peak_start_end[1] = peakEnd;
for(int sample_i=0; sample_i < n_samples; sample_i++){
peak_loss_vec[sample_i] = candidate_loss_vec[sample_i];
mean_mat[sample_i] = seg1_mean_vec[sample_i];
mean_mat[sample_i+n_samples] = seg2_mean_vec[sample_i];
mean_mat[sample_i+n_samples*2] = seg3_mean_vec[sample_i];
// Also save the left/right cumsums, which are needed to
// perform the step2 optimization.
cumsum_vec = sample_cumsum_mat + n_bins*sample_i;
if(seg1_LastIndex == 0){
left_initial_cumsum_vec[sample_i] = 0;
}else{
left_initial_cumsum_vec[sample_i] = cumsum_vec[seg1_LastIndex-1];
}
right_initial_cumsum_vec[sample_i] = cumsum_vec[seg2_LastIndex-1];
}//for(sample_i
}//if(feasible_loss < best_loss
}//if(peakEnd < unfilled_chromEnd
}//for(seg2_LastIndex
}//if(unfilled_chromStart < peakStart
}//for(seg1_LastIndex
//return status;//old end of Step1.
/* When performing the minimization over peakStart/End locations, it
* is possible that at any given bases_per_bin value, there is no
* better solution than what we found for the previous bases_per_bin
* value. In that case, we begin the search anew at a lower
* resolution, but we need to copy the cumsums from the following
* index of left_right_vec: */
int left_chromStart, right_chromStart;
/*
The while loop below corresponds to line 3 of the JointZoom
algorithm from the PeakSegJoint paper.
*/
while(1 < bases_per_bin){
best_seg1 = -1; // indicates no min found.
left_chromStart = peak_start_end[0] - bases_per_bin;
right_chromStart = peak_start_end[1] - bases_per_bin;
/*
Below in the C code we decrease the value of bases_per_bin,
as in line 4 of the JointZoom algorithm in the PeakSegJoint
paper.
*/
bases_per_bin /= bin_factor;
//printf("bases_per_bin=%d left cumsum before:\n", bases_per_bin);
for(int sample_i=0; sample_i < n_samples; sample_i++){
profile = profile_list->profile_vec + sample_i;
//printf("binSumLR sample_i=%d\n", sample_i);
status = binSumLR(data_start_end,
profile->chromStart, profile->chromEnd,
profile->coverage, profile->n_entries,
left_bin_vec, right_bin_vec,
left_chromStart, right_chromStart,
bases_per_bin, n_bins);
if(status != 0){
//printf("binSumLR bad status\n");
free(sample_count_mat);
free(last_cumsum_vec);
free(sample_cumsum_mat);
free(seg1_mean_vec);
free(seg2_mean_vec);
free(seg3_mean_vec);
free(seg1_loss_vec);
free(candidate_loss_vec);
free(left_bin_vec);
free(right_bin_vec);
free(left_cumsum_mat);
free(right_cumsum_mat);
free(left_initial_cumsum_vec);
free(right_initial_cumsum_vec);
return status;
}
left_cumsum_vec = left_cumsum_mat + n_bins*sample_i;
left_cumsum_value = left_initial_cumsum_vec[sample_i];
right_cumsum_vec = right_cumsum_mat + n_bins*sample_i;
right_cumsum_value = right_initial_cumsum_vec[sample_i];
//printf("%d ", left_cumsum_value);
//printf("%d ", right_cumsum_value);
for(int bin_i=0; bin_i < n_bins; bin_i++){
left_cumsum_value += left_bin_vec[bin_i];
left_cumsum_vec[bin_i] = left_cumsum_value;
right_cumsum_value += right_bin_vec[bin_i];
right_cumsum_vec[bin_i] = right_cumsum_value;
}
}//for(sample_i
/*
cumsum matrices have been computed, so now use them to
compute the loss and feasibility of all models.
The for loops below correspond to SearchNearPeak() on line
5 of the JointZoom algorithm in the PeakSegJoint paper.
*/
for(int seg1_LastIndex=0; seg1_LastIndex < n_bins; seg1_LastIndex++){
peakStart = left_chromStart + (seg1_LastIndex+1)*bases_per_bin;
if(unfilled_chromStart < peakStart){
//printf("[seg1last=%d] seg1 cumsum bases ", seg1_LastIndex);
for(int sample_i=0; sample_i < n_samples; sample_i++){
left_cumsum_vec = left_cumsum_mat + n_bins*sample_i;
cumsum_value = left_cumsum_vec[seg1_LastIndex];
bin_bases = peakStart - seg1_chromStart;
data_bases = bin_bases - extra_before;
seg1_mean = cumsum_value/data_bases;
//printf("%d %f ", cumsum_value, bases_value);
seg1_mean_vec[sample_i] = seg1_mean;
seg1_loss_vec[sample_i] = OptimalPoissonLoss(cumsum_value, seg1_mean);
}
//printf("\n");
for(int seg2_LastIndex=0; seg2_LastIndex < n_bins; seg2_LastIndex++){
peakEnd = right_chromStart + (seg2_LastIndex+1)*bases_per_bin;
if(peakStart < peakEnd && peakEnd < unfilled_chromEnd){
feasible_loss = 0.0;
//printf("[seg2last=%d]\n", seg2_LastIndex);
for(int sample_i=0; sample_i < n_samples; sample_i++){
left_cumsum_vec = left_cumsum_mat + n_bins*sample_i;
right_cumsum_vec = right_cumsum_mat + n_bins*sample_i;
//segment 1.
candidate_loss_vec[sample_i] = seg1_loss_vec[sample_i];
//segment 2.
cumsum_value =
right_cumsum_vec[seg2_LastIndex]-
left_cumsum_vec[seg1_LastIndex];
data_bases = peakEnd-peakStart;
seg2_mean = cumsum_value/data_bases;
seg2_mean_vec[sample_i] = seg2_mean;
candidate_loss_vec[sample_i] += OptimalPoissonLoss(cumsum_value, seg2_mean);
//segment 3.
cumsum_value =
last_cumsum_vec[sample_i]-
right_cumsum_vec[seg2_LastIndex];
bin_bases = seg3_chromEnd - peakEnd;
data_bases = bin_bases - extra_after;
seg3_mean = cumsum_value/data_bases;
seg3_mean_vec[sample_i] = seg3_mean;
candidate_loss_vec[sample_i] += OptimalPoissonLoss(cumsum_value, seg3_mean);
if(seg1_mean < seg2_mean && seg2_mean > seg3_mean){
feasible_loss += candidate_loss_vec[sample_i];
}else{
feasible_loss += flat_loss_vec[sample_i];
}
}//for(sample_i
if(feasible_loss < best_loss){
best_loss = feasible_loss;
peak_start_end[0] = peakStart;
peak_start_end[1] = peakEnd;
best_seg1 = seg1_LastIndex;
best_seg2 = seg2_LastIndex;
for(int sample_i=0; sample_i < n_samples; sample_i++){
peak_loss_vec[sample_i] = candidate_loss_vec[sample_i];
mean_mat[sample_i] = seg1_mean_vec[sample_i];
mean_mat[sample_i+n_samples] = seg2_mean_vec[sample_i];
mean_mat[sample_i+n_samples*2] = seg3_mean_vec[sample_i];
}
}//if(feasible_loss<best_loss
}//if(peakStart<peakEnd
}//for(seg2_LastIndex
}//if(unfilled_chromStart < peakStart
}//for(seg1_LastIndex
for(int sample_i=0; sample_i < n_samples; sample_i++){
left_cumsum_vec = left_cumsum_mat + n_bins*sample_i;
right_cumsum_vec = right_cumsum_mat + n_bins*sample_i;
if(best_seg1 == -1){
/* the bin_factor is the number of new bins that get put into old bins.
[ ] old bin
_ _ new bins
bin_factor=2 [ _ _ ] peakStart [ _ _ ]
bin_factor=3 [ _ _ _ ] peakStart [ _ _ _ ]
We always start the search on the new level from the old bin just
before the peakStart/End. Therefore if no new min is found, the
best choice is the old one, seg1_LastIndex=bin_factor-1 */
best_seg1 = bin_factor-1;
best_seg2 = bin_factor-1;
//Rprintf("no new min found! best_seg1=%d best_seg2=%d\n", best_seg1, best_seg2);
}
if(best_seg1 != 0){
left_initial_cumsum_vec[sample_i] = left_cumsum_vec[best_seg1-1];
}
if(best_seg2 != 0){
right_initial_cumsum_vec[sample_i] = right_cumsum_vec[best_seg2-1];
}
}
}//while(1 < bases_per_bin)
//printf("free at end\n");
free(sample_count_mat);
free(last_cumsum_vec);
free(sample_cumsum_mat);
free(seg1_mean_vec);
free(seg2_mean_vec);
free(seg3_mean_vec);
free(seg1_loss_vec);
free(candidate_loss_vec);
free(left_bin_vec);
free(right_bin_vec);
free(left_cumsum_mat);
free(right_cumsum_mat);
free(left_initial_cumsum_vec);
free(right_initial_cumsum_vec);
return 0;
}
