void AssocPairedSet::print()
{
int i;
char seq[100];
printf("print %lu elems \n",liste.size());
for (i=0; i<liste.size(); i++)
{
code2seq(liste[i], seq);
printf("%s=(", seq);
code2seq(liste_value[i].nk1.prev_kmer, seq);
printf("%s,%c) (", seq, liste_value[i].nk1.nt);
code2seq(liste_value[i].nk2.prev_kmer, seq);
printf("%s,%c)\n", seq, liste_value[i].nk2.nt);
}
}
void end_debloom_partition(bool last_partition)
{
int value;
char false_positive_kmer_char[sizeKmer+1];
FILE *file_false_positive_kmers =NULL;
kmer_type graine;
/////////////////////////begin write files
rewind (F_debloom_read);
rewind (F_debloom_write);
#ifndef MINGW
ftruncate(fileno(F_debloom_write), 0); //erase previous file
#else // tempfix? fileno is not accepted by mingw
fclose(F_debloom_write);
F_debloom_write = fopen(return_file_name("debloom2"),"wb+");
#endif
BinaryReads* file_false_positive_kmers_binary = NULL;
if (last_partition)
{
// write false positive kmers to fasta file
file_false_positive_kmers = fopen(return_file_name(false_positive_kmers_file),"wb");
char *false_positive_kmers_binary_file = (char *)"false_positive_kmers_binary";
file_false_positive_kmers_binary = new BinaryReads(
return_file_name(false_positive_kmers_binary_file), true);
//TODO: change store FP into binary? reduce space?
}
n_false_positives = 0;
while(fread(&graine, sizeof(graine),1, F_debloom_read)){
if(hasht1->get(graine,&value)==0) //kmer not present == kmer not solid
{
n_false_positives ++;
if (!fwrite(&graine, sizeof(graine), 1, F_debloom_write))
{
printf("error: can't fwrite (disk full?)\n");
exit(1);
}
if (last_partition)
{
code2seq(graine,false_positive_kmer_char);
fprintf(file_false_positive_kmers,">fp\n");
fputs(false_positive_kmer_char,file_false_positive_kmers);
fprintf(file_false_positive_kmers,"\n");
file_false_positive_kmers_binary->write_read(false_positive_kmer_char, sizeKmer);
}
}
//else kmer is a true positive, do nothing
}
if (last_partition){
fclose(file_false_positive_kmers);
file_false_positive_kmers_binary->close();
}
}
inline void assemble()
{
//////-------------------------------------------------------------------------------------------
fprintf (stderr,"______________________________________________________ \n");
fprintf (stderr,"___________ Assemble from bloom filter _______________ \n");
fprintf (stderr,"______________________________________________________ \n\n");
//////-------------------------------------------------------------------------------------------
long long len_left = 0;
long long len_right = 0;
long long contig_len =0;
long long maxlen=10000000;
char *left_traversal = (char *) malloc(maxlen*sizeof(char));
char *right_traversal = (char *) malloc(maxlen*sizeof(char));
char *contig = (char *) malloc(2*(maxlen+sizeKmer)*sizeof(char));
kmer_type kmer;
long long nbContig =0;
long long nbSmallContig =0;
long long totalnt=0;
long long max_contig_len=0;
long long mlenleft=0,mlenright=0;
int64_t NbBranchingKmer=0;
char kmer_seq[sizeKmer+1];
FILE * file_assembly = fopen(return_file_name(assembly_file),"w+");
BinaryBank *SolidKmers = new BinaryBank(return_file_name(solid_kmers_file),sizeof(kmer_type),0);
STARTWALL(assembly);
char *assemble_only_one_region = NULL; // debugging, set to a ASCII kmer to activate, NULL to desactivate
bool LOAD_BRANCHING_KMERS=false; // debugging
bool DUMP_BRANCHING_KMERS=false;
BranchingTerminator *terminator;
if (LOAD_BRANCHING_KMERS)
{
BinaryBank *BranchingKmers = new BinaryBank(return_file_name(branching_kmers_file),sizeof(kmer_type),false);
terminator = new BranchingTerminator(BranchingKmers,SolidKmers, bloo1,false_positives);
BranchingKmers->close();
}
else
terminator = new BranchingTerminator(SolidKmers,genome_size, bloo1,false_positives);
if (DUMP_BRANCHING_KMERS)
{
BinaryBank *BranchingKmers = new BinaryBank(return_file_name(branching_kmers_file),sizeof(kmer_type),true);
terminator->dump_branching_kmers(BranchingKmers);
BranchingKmers->close();
}
#ifdef UNITIG
SimplePathsTraversal *traversal = new SimplePathsTraversal(bloo1,false_positives,terminator);
fprintf (stderr,"_________________Assembling in Unitig mode ..._____________________ \n\n");
#else
MonumentTraversal *traversal = new MonumentTraversal(bloo1,false_positives,terminator);
#endif
//RandomBranchingTraversal *traversal = new RandomBranchingTraversal(bloo1,false_positives,terminator);
traversal->set_maxlen(maxlen);
traversal->set_max_depth(500);
traversal->set_max_breadth(20);
while (terminator->next(&kmer))
{
// keep looping while a starting kmer is available from this kmer
// everything will be marked during the traversal()'s
kmer_type starting_kmer;
#ifdef UNITIG
while (traversal->get_new_starting_node_improved(kmer,starting_kmer))
#else
while (traversal->find_starting_kmer(kmer,starting_kmer))
#endif
{
code2seq(starting_kmer,kmer_seq); // convert starting kmer to nucleotide seq
traversal->revert_stats(); // set stats from the last commit (discard stats from find_starting_kmer / small contigs)
if (assemble_only_one_region != NULL)
{
kmer_type dummy;
starting_kmer = extractKmerFromRead(assemble_only_one_region,0,&kmer,&dummy,false);
}
// right extension
len_right = traversal->traverse(starting_kmer,right_traversal,0);
mlenright= max(len_right,mlenright);
// left extension, is equivalent to right extension of the revcomp
len_left = traversal->traverse(starting_kmer,left_traversal,1);
mlenleft= max(len_left,mlenleft);
// form the contig
revcomp_sequence(left_traversal,len_left);
strcpy(contig,left_traversal); // contig = revcomp(left_traversal)
strcat(contig,kmer_seq);// + starting_kmer
strcat(contig,right_traversal);// + right_traversal
contig_len=len_left+len_right+sizeKmer;
// save the contig
if(contig_len >= MIN_CONTIG_SIZE)
{
max_contig_len = max(max_contig_len,contig_len);
fprintf(file_assembly,">%lli__len__%lli \n",nbContig,contig_len);
fprintf(file_assembly,"%s\n",contig);
nbContig++;
totalnt+=contig_len;
traversal->commit_stats();
}
else
{
traversal->revert_stats();
nbSmallContig++;
}
if (assemble_only_one_region != NULL)
break;
}
NbBranchingKmer++;
if ((NbBranchingKmer%300)==0) fprintf (stderr,"%cLooping through branching kmer n° %lld / %lld total nt %lld ",13,(long long int) NbBranchingKmer,(long long int) terminator->nb_branching_kmers, (long long int)totalnt );
if (nbContig > 0 && assemble_only_one_region != NULL)
break;
}
fclose(file_assembly);
fprintf (stderr,"\n Total nt assembled %lli nbContig %lli\n",totalnt,nbContig);
fprintf (stderr," Max contig len %lli (debug: max len left %lli, max len right %lli)\n",max_contig_len,mlenleft,mlenright);
fprintf (stderr,"\n Debug traversal stats: %ld ends of contigs (%lld unsaved small contigs), among them:\n",traversal->final_stats.ended_traversals,nbSmallContig);
fprintf (stderr," %ld couldn't validate consensuses\n",traversal->final_stats.couldnt_validate_consensuses);
fprintf (stderr," %ld large bubble breadth, %ld large bubble depth, %ld marked kmer, %ld no extension\n",traversal->final_stats.couldnt_traverse_bubble_breadth,traversal->final_stats.couldnt_traverse_bubble_depth,traversal->final_stats.couldnt_because_marked_kmer,traversal->final_stats.couldnt_find_extension);
fprintf (stderr," %ld in-branchin large depth, %ld in-branching large breadth, %ld in-branching other\n",traversal->final_stats.couldnt_inbranching_depth,traversal->final_stats.couldnt_inbranching_breadth,traversal->final_stats.couldnt_inbranching_other);
STOPWALL(assembly,"Assembly");
free(left_traversal);
free(right_traversal);
free(contig);
SolidKmers->close();
}