/**
* Gets a sorted string representation of a variant.
*/
void bcf_variant2string_sorted(bcf_hdr_t *h, bcf1_t *v, kstring_t *var)
{
bcf_print_liten(h,v);
bcf_unpack(v, BCF_UN_STR);
var->l = 0;
kputs(bcf_get_chrom(h, v), var);
kputc(':', var);
kputw(bcf_get_pos1(v), var);
kputc(':', var);
if (v->n_allele==2)
{
kputs(bcf_get_alt(v, 0), var);
kputc(',', var);
kputs(bcf_get_alt(v, 1), var);
}
else
{
char** allele = bcf_get_allele(v);
char** temp = (char**) malloc((bcf_get_n_allele(v)-1)*sizeof(char*));
for (int32_t i=1; i<v->n_allele; ++i)
{
temp[i] = allele[i];
}
std::qsort(temp, bcf_get_n_allele(v), sizeof(char*), cmpstr);
kputs(bcf_get_alt(v, 0), var);
for (int32_t i=0; i<v->n_allele-1; ++i)
{
kputc(',', var);
kputs(temp[i], var);
}
free(temp);
}
}
/**
* Extract reference sequence region for motif discovery in a fuzzy fashion.
*/
void CandidateRegionExtractor::extract_regions_by_fuzzy_alignment(bcf_hdr_t* h, bcf1_t* v, Variant& variant)
{
if (debug)
{
if (debug) std::cerr << "********************************************\n";
std::cerr << "EXTRACTIING REGION BY FUZZY ALIGNMENT\n\n";
}
VNTR& vntr = variant.vntr;
const char* chrom = bcf_get_chrom(h, v);
int32_t min_beg1 = bcf_get_pos1(v);
int32_t max_end1 = min_beg1;
//merge candidate search region
for (size_t i=1; i<bcf_get_n_allele(v); ++i)
{
std::string ref(bcf_get_alt(v, 0));
std::string alt(bcf_get_alt(v, i));
int32_t pos1 = bcf_get_pos1(v);
trim(pos1, ref, alt);
if (debug)
{
std::cerr << "indel fragment : " << (ref.size()<alt.size()? alt : ref) << "\n";
std::cerr << " : " << ref << ":" << alt << "\n";
}
min_beg1 = fuzzy_left_align(chrom, pos1, ref, alt, 3);
max_end1 = fuzzy_right_align(chrom, pos1 + ref.size() - 1, ref, alt, 3);
int32_t seq_len;
char* seq = faidx_fetch_seq(fai, chrom, min_beg1-1, max_end1-1, &seq_len);
if (debug)
{
std::cerr << "FUZZY REGION " << min_beg1 << "-" << max_end1 << " (" << max_end1-min_beg1+1 <<") " << "\n";
std::cerr << " " << seq << "\n";
}
if (seq_len) free(seq);
}
int32_t seq_len;
char* seq = faidx_fetch_seq(fai, chrom, min_beg1-1, max_end1-1, &seq_len);
if (debug)
{
std::cerr << "FINAL FUZZY REGION " << min_beg1 << "-" << max_end1 << " (" << max_end1-min_beg1+1 <<") " << "\n";
std::cerr << " " << seq << "\n";
}
vntr.exact_repeat_tract = seq;
vntr.exact_rbeg1 = min_beg1;
if (seq_len) free(seq);
}
/**
* Gets a string representation of a variant.
*/
void bcf_variant2string(bcf_hdr_t *h, bcf1_t *v, kstring_t *var)
{
bcf_unpack(v, BCF_UN_STR);
var->l = 0;
kputs(bcf_get_chrom(h, v), var);
kputc(':', var);
kputw(bcf_get_pos1(v), var);
kputc(':', var);
for (int32_t i=0; i<v->n_allele; ++i)
{
if (i) kputc(',', var);
kputs(bcf_get_alt(v, i), var);
}
}
/**
* Gets a string representation of the variant.
*/
std::string Variant::get_variant_string()
{
kstring_t var = {0,0,0};
bcf_unpack(v, BCF_UN_STR);
var.l = 0;
kputs(bcf_get_chrom(h, v), &var);
kputc(':', &var);
kputw(bcf_get_pos1(v), &var);
kputc(':', &var);
for (size_t i=0; i<bcf_get_n_allele(v); ++i)
{
if (i) kputc('/', &var);
kputs(bcf_get_alt(v, i), &var);
}
std::string str(var.s);
if (var.m) free(var.s);
return str;
}
/**
* Classifies variants.
*/
int32_t VariantManip::classify_variant(bcf_hdr_t *h, bcf1_t *v, Variant& var)
{
bcf_unpack(v, BCF_UN_STR);
const char* chrom = bcf_get_chrom(h, v);
uint32_t pos1 = bcf_get_pos1(v);
char** allele = bcf_get_allele(v);
int32_t n_allele = bcf_get_n_allele(v);
int32_t pos0 = pos1-1;
var.ts = 0;
var.tv = 0;
var.ins = 0;
var.del = 0;
var.clear(); // this sets the type to VT_REF by default.
bool homogeneous_length = true;
char* ref = allele[0];
int32_t rlen = strlen(ref);
//if only ref allele, skip this entire for loop
for (size_t i=1; i<n_allele; ++i)
{
int32_t type = VT_REF;
//check for tags
if (strchr(allele[i],'<'))
{
size_t len = strlen(allele[i]);
if (len>=5)
{
//VN/d+
if (allele[i][0]=='<' && allele[i][1]=='V' && allele[i][2]=='N' && allele[i][len-1]=='>' )
{
for (size_t j=3; j<len-1; ++j)
{
if (allele[i][j]<'0' || allele[i][j]>'9')
{
type = VT_VNTR;
}
}
}
//VNTR
else if (len==6 &&
allele[i][0]=='<' &&
allele[i][1]=='V' && allele[i][2]=='N' && allele[i][3]=='T' && allele[i][4]=='R' &&
allele[i][5]=='>' )
{
type = VT_VNTR;
}
//ST/d+
else if (allele[i][0]=='<' && allele[i][1]=='S' && allele[i][2]=='T' && allele[i][len-1]=='>' )
{
for (size_t j=3; j<len-1; ++j)
{
if (allele[i][j]<'0' || allele[i][j]>'9')
{
type = VT_VNTR;
}
}
}
//STR
else if (len==5 &&
allele[i][0]=='<' &&
allele[i][1]=='S' && allele[i][2]=='T' && allele[i][3]=='R' &&
allele[i][4]=='>' )
{
type = VT_VNTR;
}
}
if (type==VT_VNTR)
{
type = VT_VNTR;
var.type |= type;
var.alleles.push_back(Allele(type));
}
else
{
type = VT_SV;
var.type |= type;
std::string sv_type(allele[i]);
var.alleles.push_back(Allele(type, sv_type));
}
}
else if (allele[i][0]=='.' || strcmp(allele[i],allele[0])==0)
{
type = VT_REF;
}
else
{
kstring_t REF = {0,0,0};
kstring_t ALT = {0,0,0};
ref = allele[0];
char* alt = allele[i];
int32_t alen = strlen(alt);
if (rlen!=alen)
{
homogeneous_length = false;
}
//trimming
//this is required in particular for the
//characterization of multiallelics and
//in general, any unnormalized variant
int32_t rl = rlen;
int32_t al = alen;
//trim right
while (rl!=1 && al!=1)
{
if (ref[rl-1]==alt[al-1])
{
--rl;
--al;
}
else
{
break;
}
}
//trim left
while (rl !=1 && al!=1)
{
if (ref[0]==alt[0])
{
++ref;
++alt;
--rl;
--al;
}
else
{
break;
}
}
kputsn(ref, rl, &REF);
kputsn(alt, al, &ALT);
ref = REF.s;
alt = ALT.s;
int32_t mlen = std::min(rl, al);
int32_t dlen = al-rl;
int32_t diff = 0;
int32_t ts = 0;
int32_t tv = 0;
if (mlen==1 && dlen)
{
char ls, le, ss;
if (rl>al)
{
ls = ref[0];
le = ref[rl-1];
ss = alt[0];
}
else
{
ls = alt[0];
le = alt[al-1];
ss = ref[0];
}
if (ls!=ss && le!=ss)
{
++diff;
if ((ls=='G' && ss=='A') ||
(ls=='A' && ss=='G') ||
(ls=='C' && ss=='T') ||
(ls=='T' && ss=='C'))
{
++ts;
}
else
{
++tv;
}
}
}
else
{
for (int32_t j=0; j<mlen; ++j)
{
if (ref[j]!=alt[j])
{
++diff;
if ((ref[j]=='G' && alt[j]=='A') ||
(ref[j]=='A' && alt[j]=='G') ||
(ref[j]=='C' && alt[j]=='T') ||
(ref[j]=='T' && alt[j]=='C'))
{
++ts;
}
else
{
++tv;
}
}
}
}
//substitution variants
if (mlen==diff)
{
type |= mlen==1 ? VT_SNP : VT_MNP;
}
//indel variants
if (dlen)
{
type |= VT_INDEL;
}
//clumped SNPs and MNPs
if (diff && diff < mlen) //internal gaps
{
type |= VT_CLUMPED;
}
var.type |= type;
var.alleles.push_back(Allele(type, diff, alen, dlen, mlen, ts, tv));
var.ts += ts;
var.tv += tv;
var.ins = dlen>0?1:0;
var.del = dlen<0?1:0;
if (REF.m) free(REF.s);
if (ALT.m) free(ALT.s);
}
}
if (var.type==VT_VNTR)
{
bcf_unpack(v, BCF_UN_INFO);
//populate motif, motif len etc. etc.
// char* str = NULL;
// int32_t n = 0;
// int32_t ret = bcf_get_info_string(h, v, "MOTIF", &str, &n);
// if (ret>0)
// {
// var.motif = std::string(str);
// var.mlen = var.motif.size();
// }
// ret = bcf_get_info_string(h, v, "RU", &str, &n);
// if (ret>0)
// {
// var.ru = std::string(str);
// var.mlen = var.ru.size();
// }
// if (n) free(str);
//
// int32_t* no = NULL;
// n = 0;
// ret = bcf_get_info_int32(h, v, "RL", &no, &n);
// if (ret>0) var.rlen = *no;
// if (n) free(no);
//
// int32_t* fl = NULL;
// n = 0;
// ret = bcf_get_info_int32(h, v, "REF", &fl, &n);
// if (ret>0) var.rcn = *fl;
// if (n) free(fl);
}
//additionally define MNPs by length of all alleles
if (!(var.type&(VT_VNTR|VT_SV)) && var.type!=VT_REF)
{
if (homogeneous_length && rlen>1 && n_allele>1)
{
var.type |= VT_MNP;
}
}
return var.type;
}
/**
* Detects near by STRs.
*/
bool VariantManip::detect_str(bcf_hdr_t *h, bcf1_t *v, Variant& variant)
{
return detect_str(bcf_get_chrom(h, v), bcf_get_pos1(v), variant);
}
/**
* Extract reference sequence region for motif discovery.
*
* The input is a VCF record that contains an indel.
*
* If the the indel has multiple alleles, it will examine all
* alleles.
*
* todo: is might be a good idea to combine this step with motif detection
* since there seems to be a need to have an iterative process here
* to ensure a good candidate motif is chosen. *
*/
void CandidateRegionExtractor::extract_regions_by_exact_alignment(bcf_hdr_t* h, bcf1_t* v, Variant& variant)
{
if (debug)
{
if (debug) std::cerr << "********************************************\n";
std::cerr << "EXTRACTIING REGION BY EXACT LEFT AND RIGHT ALIGNMENT\n\n";
}
VNTR& vntr = variant.vntr;
const char* chrom = bcf_get_chrom(h, v);
int32_t min_beg1 = bcf_get_pos1(v);
int32_t max_end1 = min_beg1;
if (debug)
{
bcf_print_liten(h, v);
}
//merge candidate search region
for (size_t i=1; i<bcf_get_n_allele(v); ++i)
{
std::string ref(bcf_get_alt(v, 0));
std::string alt(bcf_get_alt(v, i));
int32_t pos1 = bcf_get_pos1(v);
//this prevents introduction of flanks that do not harbour the repeat unit
trim(pos1, ref, alt);
int32_t end1 = pos1 + ref.size() - 1;
right_align(chrom, end1, ref, alt);
int32_t beg1 = end1 - ref.size() + 1;
left_align(chrom, beg1, ref, alt);
min_beg1 = beg1<min_beg1 ? beg1 : min_beg1;
max_end1 = end1>max_end1 ? end1 : max_end1;
int32_t seq_len;
char* seq = faidx_fetch_seq(fai, chrom, min_beg1-1, max_end1-1, &seq_len);
if (debug)
{
std::cerr << "EXACT REGION " << min_beg1 << "-" << max_end1 << " (" << max_end1-min_beg1+1 <<") from " << pos1 << ":" << ref << ":" << alt << "\n";
std::cerr << " " << seq << "\n";
}
if (seq_len) free(seq);
}
int32_t seq_len;
char* seq = faidx_fetch_seq(fai, chrom, min_beg1-1, max_end1-1, &seq_len);
if (debug)
{
std::cerr << "FINAL EXACT REGION " << min_beg1 << "-" << max_end1 << " (" << max_end1-min_beg1+1 <<") " << "\n";
std::cerr << " " << seq << "\n";
}
vntr.exact_repeat_tract = seq;
vntr.rid = bcf_get_rid(v);
vntr.exact_rbeg1 = min_beg1;
vntr.exact_rend1 = max_end1;
if (seq_len) free(seq);
}
/**
* Gets sequence name of a record.
*/
const char* BCFSyncedStreamReader::get_seqname(int32_t i, bcf1_t *v)
{
return bcf_get_chrom(hdrs[i], v);
}
/**
* Constructor.
*/
Variant::Variant(bcf_hdr_t* h, bcf1_t* v)
{
this->h = h;
this->v = v;
type = classify(h, v);
chrom = bcf_get_chrom(h, v);
rid = bcf_get_rid(v);
pos1 = bcf_get_pos1(v);
no_overlapping_snps = 0;
no_overlapping_indels = 0;
no_overlapping_vntrs = 0;
is_new_multiallelic = false;
//attempts to update relevant information on variants
if (type==VT_SNP)
{
beg1 = bcf_get_pos1(v);
end1 = bcf_get_pos1(v);
}
else if (type==VT_INDEL)
{
beg1 = bcf_get_pos1(v);
end1 = bcf_get_info_int(h, v, "END", 0);
//annotate ends
if (!end1) end1 = bcf_get_end1(v);
}
//complex variants
else if (type & (VT_SNP|VT_MNP|VT_INDEL|VT_CLUMPED))
{
beg1 = bcf_get_pos1(v);
end1 = bcf_get_info_int(h, v, "END", 0);
if (!end1) end1 = bcf_get_end1(v);
}
else if (type==VT_VNTR)
{
beg1 = bcf_get_pos1(v);
end1 = bcf_get_info_int(h, v, "END", 0);
if (!end1) end1 = bcf_get_end1(v);
update_vntr_from_info_fields(h, v);
vs.push_back(v);
vntr_vs.push_back(v);
}
else if (type==VT_SV)
{
beg1 = bcf_get_pos1(v);
end1 = bcf_get_info_int(h, v, "END", 0);
if (!end1) end1 = bcf_get_end1(v);
}
else
{
std::cerr << "unexpected type in variant construction\n";
print();
exit(1);
}
}
/**
* Classifies variants.
*/
int32_t Variant::classify(bcf_hdr_t *h, bcf1_t *v)
{
clear();
this->h = h;
this->v = v;
bcf_unpack(v, BCF_UN_STR);
chrom.assign(bcf_get_chrom(h, v));
rid = bcf_get_rid(v);
pos1 = bcf_get_pos1(v);
end1 = bcf_get_end1(v);
char** allele = bcf_get_allele(v);
int32_t n_allele = bcf_get_n_allele(v);
int32_t pos0 = pos1-1;
bool homogeneous_length = true;
char* ref = allele[0];
int32_t rlen = strlen(ref);
if (strchr(ref, 'N'))
{
contains_N = true;
}
//if only ref allele, skip this entire for loop
for (size_t i=1; i<n_allele; ++i)
{
int32_t allele_type = VT_REF;
//check for symbolic alternative alleles
if (strchr(allele[i],'<'))
{
size_t len = strlen(allele[i]);
if (len>=5)
{
//VN/d+
if (allele[i][0]=='<' && allele[i][1]=='V' && allele[i][2]=='N' && allele[i][len-1]=='>' )
{
for (size_t j=3; j<len-1; ++j)
{
if (allele[i][j]<'0' || allele[i][j]>'9')
{
allele_type = VT_VNTR;
}
}
}
//VNTR
else if (len==6 &&
allele[i][0]=='<' &&
allele[i][1]=='V' && allele[i][2]=='N' && allele[i][3]=='T' && allele[i][4]=='R' &&
allele[i][5]=='>' )
{
allele_type = VT_VNTR;
}
//STR
else if (len==5 &&
allele[i][0]=='<' &&
allele[i][1]=='S' && allele[i][2]=='T' && allele[i][3]=='R' &&
allele[i][4]=='>' )
{
allele_type = VT_VNTR;
}
//ST/d+
else if (allele[i][0]=='<' && allele[i][1]=='S' && allele[i][2]=='T' && allele[i][len-1]=='>' )
{
type = VT_VNTR;
for (size_t j=3; j<len-1; ++j)
{
if ((allele[i][j]<'0' || allele[i][j]>'9') && allele[i][j]!='.')
{
type = VT_SV;
}
}
}
}
if (allele_type==VT_VNTR)
{
allele_type = VT_VNTR;
type |= allele_type;
alleles.push_back(Allele(allele_type));
}
else
{
allele_type = VT_SV;
type |= allele_type;
std::string sv_type(allele[i]);
alleles.push_back(Allele(allele_type, sv_type));
}
}
//checks for chromosomal breakpoints
else if (strchr(allele[i],'[')||strchr(allele[i],']'))
{
allele_type = VT_SV;
type |= allele_type;
std::string sv_type("<BND>");
alleles.push_back(Allele(allele_type, sv_type));
}
//non variant record
else if (allele[i][0]=='.' || strcmp(allele[i],allele[0])==0)
{
type = VT_REF;
}
//explicit sequence of bases
else
{
kstring_t REF = {0,0,0};
kstring_t ALT = {0,0,0};
ref = allele[0];
char* alt = allele[i];
int32_t alen = strlen(alt);
if (strchr(alt, 'N'))
{
contains_N = true;
}
if (rlen!=alen)
{
homogeneous_length = false;
}
//trimming
//this is required in particular for the
//characterization of multiallelics and
//in general, any unnormalized variant
int32_t rl = rlen;
int32_t al = alen;
//trim right
while (rl!=1 && al!=1)
{
if (ref[rl-1]==alt[al-1])
{
--rl;
--al;
}
else
{
break;
}
}
//trim left
while (rl !=1 && al!=1)
{
if (ref[0]==alt[0])
{
++ref;
++alt;
--rl;
--al;
}
else
{
break;
}
}
kputsn(ref, rl, &REF);
kputsn(alt, al, &ALT);
ref = REF.s;
alt = ALT.s;
int32_t mlen = std::min(rl, al);
int32_t dlen = al-rl;
int32_t diff = 0;
int32_t ts = 0;
int32_t tv = 0;
if (mlen==1 && dlen)
{
char ls, le, ss;
if (rl>al)
{
ls = ref[0];
le = ref[rl-1];
ss = alt[0];
}
else
{
ls = alt[0];
le = alt[al-1];
ss = ref[0];
}
if (ls!=ss && le!=ss)
{
++diff;
if ((ls=='G' && ss=='A') ||
(ls=='A' && ss=='G') ||
(ls=='C' && ss=='T') ||
(ls=='T' && ss=='C'))
{
++ts;
}
else
{
++tv;
}
}
}
else
{
for (int32_t j=0; j<mlen; ++j)
{
if (ref[j]!=alt[j])
{
++diff;
if ((ref[j]=='G' && alt[j]=='A') ||
(ref[j]=='A' && alt[j]=='G') ||
(ref[j]=='C' && alt[j]=='T') ||
(ref[j]=='T' && alt[j]=='C'))
{
++ts;
}
else
{
++tv;
}
}
}
}
//substitution variants
if (mlen==diff)
{
allele_type |= mlen==1 ? VT_SNP : VT_MNP;
}
//indel variants
if (dlen)
{
allele_type |= VT_INDEL;
}
//clumped SNPs and MNPs
if (diff && diff < mlen) //internal gaps
{
allele_type |= VT_CLUMPED;
}
type |= allele_type;
alleles.push_back(Allele(type, diff, alen, dlen, mlen, ts, tv));
ts += ts;
tv += tv;
ins = dlen>0?1:0;
del = dlen<0?1:0;
if (REF.m) free(REF.s);
if (ALT.m) free(ALT.s);
}
}
if (type==VT_VNTR)
{
update_vntr_from_info_fields(h, v);
}
//additionally define MNPs by length of all alleles
if (!(type&(VT_VNTR|VT_SV)) && type!=VT_REF)
{
if (homogeneous_length && rlen>1 && n_allele>1)
{
type |= VT_MNP;
}
}
return type;
}
/**
* Gets sequence name of a record.
*/
const char* BCFOrderedReader::get_seqname(bcf1_t *v)
{
return bcf_get_chrom(hdr, v);
};