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));
KmerColour *left_colour_traversal = (KmerColour *) malloc(maxlen*sizeof(KmerColour));
KmerColour *right_colour_traversal = (KmerColour *) malloc(maxlen*sizeof(KmerColour));
char *contig = (char *) malloc(2*(maxlen+sizeKmer)*sizeof(char));
KmerColour *contig_colour = (KmerColour *) malloc(2*(maxlen+sizeKmer)*sizeof(KmerColour));
kmer_type kmer;
long long nbContig =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+");
FILE * file_colour_assembly = fopen(return_file_name(assembly_colour_file),"w+");
BinaryBank *SolidKmers = new BinaryBank(return_file_name(solid_kmers_file),sizeof(kmer_type),0);
BinaryBank *solid_kmers_colour = new BinaryBank(return_file_name(solid_kmers_colour_file), kSizeOfKmerType+kSizeOfKmerColour, 0);
char colour_seq[1000];
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)
{printf("LOA:%d\n",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)
{printf("DUMP:%d\n",DUMP_BRANCHING_KMERS);
BinaryBank *BranchingKmers = new BinaryBank(return_file_name(branching_kmers_file),sizeof(kmer_type),true);
terminator->dump_branching_kmers(BranchingKmers);
BranchingKmers->close();
}
printf("Check boolean:%i\t%i\n", LOAD_BRANCHING_KMERS, DUMP_BRANCHING_KMERS);
#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);
traversal->SetSolidKmersColour(solid_kmers_colour, max_memory);
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;
code2seq(kmer,kmer_seq); // convert
// printf("StartWhile, init Kmer:%li\t%s\n",kmer, kmer_seq);// Varified! kmer's matched seq from the original creation
while (traversal->find_starting_kmer(kmer,starting_kmer))
// while (traversal->find_starting_kmer_inside_simple_path(kmer,starting_kmer))
{
code2seq(starting_kmer,kmer_seq); // convert starting kmer to nucleotide seq
KmerColour kmer_colour = traversal->GetColour(starting_kmer);
// printf("Starting_kmer:%lu %s",starting_kmer, kmer_seq);
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);
len_right = traversal->traverse_colour(starting_kmer, right_traversal, right_colour_traversal, 0);
mlenright= max(len_right,mlenright);
int debug=1;
if(debug>1){
printf("RightSeq:%lld\t%s\n", len_right, right_traversal);
// printf("RightColour:");
// for (int i = 0; i < len_right; ++i) {
// printf("%u ",right_colour_traversal[i]);
// }
kmer_colour_pattern_string(right_colour_traversal, len_right, colour_seq);
printf("RightColour:%s\n",colour_seq);
}
// left extension, is equivalent to right extension of the revcomp
// len_left = traversal->traverse(starting_kmer, left_traversal, 1);
len_left = traversal->traverse_colour(starting_kmer, left_traversal,
left_colour_traversal, 1);
mlenleft= max(len_left,mlenleft);
// form the contig
// printf("before Rev:%s\n",left_traversal);
revcomp_sequence(left_traversal,len_left);
KmerColourUtil::rev_colour(left_colour_traversal, len_left);
// printf("after Rev:%s\n",left_traversal);
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;
int colour_len = 0;
KmerColour sep_colour = kErrorCode+1;// output with %x, so anything greater than 100;
colour_len = KmerColourUtil::append_colour(left_colour_traversal, len_left,
contig_colour, colour_len);
if(debug){
KmerColourUtil::append_colour(&sep_colour, 1, contig_colour,
colour_len);
}
// memset(contig_colour+pt_len, (int) kmer_colour, kSizeOfKmerColour*sizeKmer);
// pt_len += sizeKmer;
KmerColourUtil::append_colour(&kmer_colour, 1, contig_colour,
colour_len);
if(debug){
KmerColourUtil::append_colour(&sep_colour, 1, contig_colour,
colour_len);
}
// memcpy(contig_colour+colour_len, right_colour_traversal, len_right);
// colour_len += len_right;
KmerColourUtil::append_colour(right_colour_traversal, len_right,
contig_colour, colour_len);
if(debug>1){
printf("LeftSeq:%lld\t%s\n", len_left, left_traversal);
// printf("LeftColour:");
// for (int i = 0; i < len_left; ++i) {
// printf("%u ",left_colour_traversal[i]);
// }
// printf("\n");
kmer_colour_pattern_string(left_colour_traversal, len_left, colour_seq);
printf("LeftColour:%s\n",colour_seq);
printf("Kmer:%s\n",kmer_seq);
printf("KmerColour:%u\n",kmer_colour);
printf("Contig:%lld\t%s\n",contig_len ,contig);
// printf("Colour:");
// for (int i = 0; i < pt_len; ++i) {
// printf("%x", contig_colour[i]);
// }
// printf("Colour:%d\t%s\n\n",pt_len+len_right ,contig_colour);
}
std::string report("==========Summary==========\n");
// KmerColourUtil::summary(report, contig_colour, colour_len);
// KmerColourUtil::colour_table(report, contig_colour, colour_len, max_colour_count);
// printf("%s", report.data());
KmerColourSummary kcs(contig_colour, colour_len, max_colour_count);
kcs.summary_colour_code(report);
kcs.summary_colour_count(report);
kcs.summary_stat(report);
kcs.colour_table(report);
// printf("%s", report.data());
// delete &kcs;
// delete &kcs;
// printf("================END======================\n\n\n");
// save the contig
if(contig_len >= MIN_CONTIG_SIZE)//TODO: add colour info here
{
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);
fprintf(file_colour_assembly,">%lli__len__%lli \n",nbContig,contig_len);
fprintf(file_colour_assembly,"%s\n",contig);
//// fprintf(file_colour_assembly,"%s\n",contig_colour);
// for (int i = 0; i < colour_len; ++i) {
// fprintf(file_colour_assembly, "%d", all_colour[i]);
// }
fprintf(file_colour_assembly,"%s\n",report.data());
nbContig++;
totalnt+=contig_len;
}
if (assemble_only_one_region != NULL)
break;
// printf("Memory: %zu %zu\n", getPeakRSS(), getCurrentRSS() );
//exit(-1);
}
// printf("Done while look is assemble()\n");
//fclose(file_assembly);
//fclose(file_colour_assembly);
//exit(-2);
NbBranchingKmer++;
if ((NbBranchingKmer%300)==0) fprintf (stderr,"%cLooping through branching kmer n° %" PRId64 "/ %" PRId64 " total nt %lli" ,13,NbBranchingKmer,terminator->nb_branching_kmers,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,"\n Max contig len %lli (debug: max len left %lli, max len right %lli)\n",max_contig_len,mlenleft,mlenright);
STOPWALL(assembly,"Assembly");
free(left_traversal);
free(right_traversal);
free(contig);
SolidKmers->close();
solid_kmers_colour->close();
// delete SolidKmers;
// delete solid_kmers_colour;
// delete terminator;
delete traversal;
// printf("Memory: %zu %zu\n", getPeakRSS(), getCurrentRSS() );
printf("===========DONE=========EXIT========\n");
//exit(-9);
}
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();
}