// exact extend
// leav_len: leaving length to extend
// type: 1 - backward, 0 - forward
int bwt_extend_exact(const Idx2BWT *bi_bwt, ubyte_t *seq, int len, int *leav_len, int type,
bwtint_t *sa_begin, bwtint_t *sa_end, bwtint_t *rev_sa_begin, bwtint_t *rev_sa_end)
{
bwtint_t k, l, rev_k, rev_l;
int i;
k = *sa_begin;
l = *sa_end;
rev_k = *rev_sa_begin;
rev_l = *rev_sa_end;
if(type == 1){
for(i = *leav_len; i > 0; --i){
if(seq[i - 1] > 3)
break;
BWTSARangeBackward_Bidirection(bi_bwt, seq[i], &k, &l, &rev_k, &rev_l);
if(k > l)
break;
else{
*sa_begin = k;
*sa_end = l;
*rev_sa_begin = rev_k;
*rev_sa_end = rev_l;
}
}
} else {
for(i = (*leav_len); i > 0; --i){
if(seq[len-i] > 3)
break;
BWTSARangeForward_Bidirection(bi_bwt, seq[len-i], &k, &l, &rev_k, &rev_l);
if(k > l)
break;
else{
*sa_begin = k;
*sa_end = l;
*rev_sa_begin = rev_k;
*rev_sa_end = rev_l;
}
}
}
*leav_len = i;
return *sa_end - (*sa_begin) + 1;
}
int main() {
int i,j,k,c;
//Variables for backward and forward search
unsigned int l,r,rev_l,rev_r;
//Variables for search all sa ranges functions
unsigned int result_l[ALPHABET_SIZE];
unsigned int result_r[ALPHABET_SIZE];
unsigned int result_rev_l[ALPHABET_SIZE];
unsigned int result_rev_r[ALPHABET_SIZE];
//Variables for result
unsigned int offset;
int sequenceId;
unsigned int saCount;
//Variables for pattern
char pattern[1024];
strcpy(pattern,"AACCCTAACCCTAACCCTAACCCTAACCCTAACCCTAACCCTAACCCTAACCCTAACCCTAACCCTAA");
int patternLength = strlen(pattern);
// Load up the index with the below statement
printf("Loading index ... ");
fflush(stdout);
Idx2BWT * idx2BWT = BWTLoad2BWT("ncbi.genome.fa.index",".sa");
printf("DONE\n\n");
// Convert the pattern into 2BWT recognised coding scheme
unsigned char packedPattern[1024];
BWTConvertPattern(idx2BWT,pattern,patternLength,packedPattern);
// The following performs a backward search of the pattern
// ===================================================================================
// |
printf("Performing backward search of the pattern..\n");
BWTSARangeInitial(idx2BWT,packedPattern[patternLength-1],&l,&r);
for (i=patternLength-2;i>=0;i--) {
BWTSARangeBackward(idx2BWT,packedPattern[i],&l,&r);
}
printf("SA Range being = %u %u (%u)\n\n",l,r,r-l+1);
// |
// ===================================================================================
// The following performs a forward search of the pattern
// ===================================================================================
// |
printf("Performing forward search of the pattern..\n");
BWTSARangeInitial(idx2BWT,packedPattern[0],&l,&r);
BWTSARangeInitial(idx2BWT,packedPattern[0],&rev_l,&rev_r);
for (i=1;i<patternLength;i++) {
BWTSARangeForward_Bidirection(idx2BWT,packedPattern[i],&l,&r,&rev_l,&rev_r);
}
printf("SA Range being = %u %u %u %u (%u)\n\n",l,r,rev_l,rev_r,r-l+1);
// |
// ===================================================================================
// The following performs a bi-directional search of the pattern
// Starting from the middle of the pattern, first move right, then move left.
// ===================================================================================
// |
printf("Performing bi-directional search of the pattern..\n");
j = patternLength / 2;
BWTSARangeInitial(idx2BWT,packedPattern[j],&l,&r);
BWTSARangeInitial(idx2BWT,packedPattern[j],&rev_l,&rev_r);
for (i=j+1;i<patternLength;i++) {
BWTSARangeForward_Bidirection(idx2BWT,packedPattern[i],&l,&r,&rev_l,&rev_r);
}
for (i=j-1;i>=0;i--) {
BWTSARangeBackward_Bidirection(idx2BWT,packedPattern[i],&l,&r,&rev_l,&rev_r);
}
printf("SA Range being = %u %u %u %u (%u)\n\n",l,r,rev_l,rev_r,r-l+1);
// |
// ===================================================================================
// The following performs a 1-mismatch search of the pattern
// ===================================================================================
// |
// |
printf("Performing 1-mismatch search of the pattern..\n");
saCount = 0;
j = patternLength / 2;
BWTSARangeInitial(idx2BWT,packedPattern[patternLength-1],&l,&r);
for (i=patternLength-2;i>j-1;i--) { BWTSARangeBackward(idx2BWT,packedPattern[i],&l,&r); }
for (i=j-1;i>=0;i--) {
BWTAllSARangesBackward(idx2BWT,l,r,result_l,result_r);
for (c=0;c<ALPHABET_SIZE;c++) {
if (c==packedPattern[i]) continue;
unsigned int err_l=result_l[c];
unsigned int err_r=result_r[c];
for (k=i-1;k>=0;k--) {
if (err_l>err_r) break;
BWTSARangeBackward(idx2BWT,packedPattern[k],&err_l,&err_r);
}
if (err_l<=err_r && k<0) {
//An SA range of occurrence is found (err_l,err_r)
saCount+=err_r-err_l+1;
}
}
l=result_l[packedPattern[i]];
r=result_r[packedPattern[i]];
}
BWTSARangeInitial(idx2BWT,packedPattern[0],&l,&r);
BWTSARangeInitial(idx2BWT,packedPattern[0],&rev_l,&rev_r);
for (i=1;i<j;i++) { BWTSARangeForward_Bidirection(idx2BWT,packedPattern[i],&l,&r,&rev_l,&rev_r); }
for (i=j;i<patternLength;i++) {
BWTAllSARangesForward_Bidirection(idx2BWT,l,r,rev_l,rev_r,result_l,result_r,result_rev_l,result_rev_r);
for (c=0;c<ALPHABET_SIZE;c++) {
if (c==packedPattern[i]) continue;
unsigned int err_l=result_l[c];
unsigned int err_r=result_r[c];
unsigned int rev_err_l=result_rev_l[c];
unsigned int rev_err_r=result_rev_r[c];
for (k=i+1;k<patternLength;k++) {
if (err_l>err_r) break;
BWTSARangeForward_Bidirection(idx2BWT,packedPattern[k],&err_l,&err_r,&rev_err_l,&rev_err_r);
}
if (err_l<=err_r && k>=patternLength) {
//An SA range of occurrence is found (err_l,err_r)
saCount+=err_r-err_l+1;
}
}
l=result_l[packedPattern[i]];
r=result_r[packedPattern[i]];
rev_l=result_rev_l[packedPattern[i]];
rev_r=result_rev_r[packedPattern[i]];
}
printf("%u SA-indexes/occurrences were found.\n\n",saCount);
// |
// |
// ===================================================================================
// The following output the first 5 position of the pattern
// ===================================================================================
// |
// |
j=(r-l+1<5)?r-l+1:5;
printf("Reporting %d arbitrary occurrences..\n",j);
for (i=0;i<j;i++) {
BWTRetrievePositionFromSAIndex(idx2BWT,l+i,&sequenceId,&offset);
printf("Occurrence found in sequence #%d with offset %u\n",sequenceId,offset);
}
// |
// |
// ===================================================================================
// Free up the 2BWT index
printf("\nFree index ... ");
fflush(stdout);
BWTFree2BWT(idx2BWT);
printf("DONE\n");
return 0;
}