bool SeqPath::check_if_correct(const string& str, const Config *config) const
{
const vector<mass_t>& aa2mass = config->get_aa2mass();
const char *path_str = seq_str.c_str();
const char *corr_str = str.c_str();
int len_path_str = strlen(path_str);
int len_corr_str = strlen(corr_str);
if (len_path_str>len_corr_str)
return false;
int i;
for (i=0; i<=len_corr_str-len_path_str; i++)
{
int j;
bool correct_seq = true;
for (j=0; j<len_path_str; j++)
if (! (path_str[j] == corr_str[i+j] ||
(path_str[j] == 'I' && corr_str[i+j]== 'L') ||
(path_str[j] == 'L' && corr_str[i+j]== 'I') ||
(path_str[j] == 'Q' && corr_str[i+j]== 'K') ||
(path_str[j] == 'K' && corr_str[i+j]== 'Q') ) )
{
correct_seq = false;
break;
}
if (correct_seq)
{
// check prefix mass
Peptide pep;
pep.parseFromString(config,corr_str);
const vector<int>& aas= pep.get_amino_acids();
mass_t mass=0;
int j;
if (n_term_mass == 0 && i==0)
return true;
for (j=0; j<aas.size(); j++)
{
mass+=aa2mass[aas[j]];
if (fabs(mass-this->n_term_mass)<6)
return true;
if (mass>n_term_mass)
break;
}
}
}
return false;
}
int SeqPath::get_num_correct_aas(const Peptide& pep, const Config *config) const
{
const vector<mass_t>& aa2mass = config->get_aa2mass();
const vector<int>& pep_aas = pep.get_amino_acids();
int num_correct=0;
int i;
vector<mass_t> pep_masses;
vector<int> path_aas;
get_amino_acids(path_aas);
pep_masses.resize(pep_aas.size(),0);
for (i=1; i<pep_aas.size(); i++)
pep_masses[i]=pep_masses[i-1]+aa2mass[pep_aas[i-1]];
mass_t path_mass = n_term_mass;
for (i=0; i<path_aas.size(); i++)
{
const int path_aa = path_aas[i];
int j;
for (j=0; j<pep_aas.size(); j++)
{
const int pep_aa = pep_aas[j];
if (fabs(pep_masses[j]-path_mass)<1.0 && pep_aas[j] == path_aas[i])
{
num_correct++;
break;
}
}
path_mass += aa2mass[path_aas[i]];
}
return num_correct;
}
int Peptide::calc_number_of_correct_aas(Config *config,const Peptide& other) const
{
const vector<int>& other_amino_acids = other.get_amino_acids();
const int num_aas = amino_acids.size();
const int num_other_aas = other_amino_acids.size();
vector<mass_t> this_breakages,other_breakages;
calc_expected_breakage_masses(config,this_breakages);
other.calc_expected_breakage_masses(config,other_breakages);
int this_idx=0;
int other_idx=0;
int num_correct_aas=0;
while (this_idx<num_aas && other_idx<num_other_aas)
{
if (fabs(this_breakages[this_idx]-other_breakages[other_idx])<1.0)
{
if (amino_acids[this_idx]==other_amino_acids[other_idx])
num_correct_aas++;
this_idx++;
other_idx++;
}
else
{
if (this_breakages[this_idx]<other_breakages[other_idx])
{
this_idx++;
}
else
other_idx++;
}
}
return num_correct_aas;
}
void AdvancedScoreModel::score_peptide_node_combos(PrmGraph *prm, const Peptide& peptide ) const
{
const vector<int>& org_aas = config.get_org_aa();
const vector<mass_t>& aa2mass = config.get_aa2mass();
const vector<MultiEdge>& multi_edges = prm->get_multi_edges();
const int num_nodes = prm->get_num_nodes();
const vector<int>& pep_aas = peptide.get_amino_acids();
const int num_pep_aas = pep_aas.size();
mass_t p_mass=0;
int aa_idx=0;
int i;
for (i=0; i<num_nodes; i++)
{
Node& node = prm->get_non_const_node(i);
const RegionalScoreModel& score_model =
regional_breakage_score_models[prm->get_charge()][prm->get_size_idx()][node.breakage.region_idx];
int in_edge_idx=NEG_INF, in_edge_variant=NEG_INF;
int out_edge_idx=NEG_INF, out_edge_variant=NEG_INF;
// cout << "N: " << node.mass << endl;
while (aa_idx<pep_aas.size() && fabs(p_mass-node.mass)>0.1)
{
p_mass += aa2mass[pep_aas[aa_idx]];
aa_idx++;
// cout << aa_idx << "\t" << p_mass << endl;
}
if (aa_idx == pep_aas.size() && i != num_nodes-1)
{
int j;
for (j=0; j<num_nodes; j++)
cout << j << "\t" << prm->get_node(j).mass << endl;
cout << endl << "PEP:" << endl;
vector<mass_t> exp_masses;
peptide.calc_expected_breakage_masses((Config *)&config,exp_masses);
for (j=0; j<exp_masses.size(); j++)
cout << j << "\t" << exp_masses[j] << endl;
cout << "Error: mismatch between nodes and peptide!" << endl;
exit(1);
}
if (node.in_edge_idxs.size()>0)
{
int j;
for (j=0; j<node.in_edge_idxs.size(); j++)
{
const int edge_idx = node.in_edge_idxs[j];
const MultiEdge& in_edge = multi_edges[edge_idx];
const int num_aa = in_edge.num_aa;
if (num_aa>aa_idx)
continue;
const int var_idx = in_edge.get_variant_idx(num_aa,&pep_aas[aa_idx-num_aa]);
if (var_idx<0)
continue;
in_edge_idx = edge_idx;
in_edge_variant = var_idx;
break;
}
}
if (node.out_edge_idxs.size()>0)
{
int j;
for (j=0; j<node.out_edge_idxs.size(); j++)
{
const int edge_idx = node.out_edge_idxs[j];
const MultiEdge& out_edge = multi_edges[edge_idx];
const int num_aa = out_edge.num_aa;
if (num_aa + aa_idx >num_pep_aas)
continue;
const int var_idx = out_edge.get_variant_idx(num_aa,&pep_aas[aa_idx]);
if (var_idx<0)
continue;
out_edge_idx = edge_idx;
out_edge_variant = var_idx;
break;
}
}
BreakageInfo info;
prm->fill_breakage_info(this,&info,i,in_edge_idx,in_edge_variant,out_edge_idx,out_edge_variant);
node.score_combos.clear();
// cout << in_edge_idx << " " << in_edge_variant << " " << out_edge_idx << " " << out_edge_variant << "\t";
info.score = score_model.score_a_single_breakage_combo(prm, node, &node.breakage, info);
node.score_combos[ScoreComboLoc(info)]=info.score;
node.score = info.score;
node.breakage.score = node.score;
// cout << node.score << endl;
}
prm->set_has_node_combo_scores(true);
}
/**************************************************************************
Returns the global edit distance between two peptides.
Gap cost = 1.
d(I,L)=0
d(K,Q)=0
d(F/M*)=0
d(N,D)=0.5;
d(X,X)=0
d(X,Y)=1
***************************************************************************/
float Peptide::peptide_edit_distance(Config *config, Peptide& other_pep) const
{
int i;
vector<int> other_aa = other_pep.get_amino_acids();
const int *pep1 = &amino_acids[0];
const int pep_len1 = amino_acids.size();
const int *pep2 = &other_aa[0];
const int pep_len2 = other_aa.size();
const int max_width = 5;
const int ox_met_aa = config->get_aa_from_label(string("M+16"));
float row1[max_width*2+1], row2[max_width*2+1];
float *old_row, *new_row;
if (abs(pep_len1-pep_len2)>=max_width)
return pep_len1;
// check that no little switch of two aa can make them the same
if (pep_len1 == pep_len2)
{
int i;
int err_pos=-1;
int errs=0;
for (i=0; i<pep_len1; i++)
{
if ( (pep1[i] != pep2[i]) &&
! (((pep1[i]==Ile ||pep1[i]==Leu) && (pep2[i]==Ile || pep2[i]==Leu)) ||
((pep1[i]==Gln || pep1[i]==Lys) && (pep2[i]==Gln || pep2[i]==Lys)) ) )
{
err_pos=i;
errs++;
}
if (errs>2)
break;
}
if (errs == 0)
return 0;
if (errs == 2)
{
if ( (pep1[err_pos] == pep2[err_pos-1]) && (pep1[err_pos-1] == pep2[err_pos]))
return 1;
}
}
old_row=row1;
new_row=row2;
for (i=0; i<max_width; i++)
{
old_row[i]=9999;
old_row[i+max_width]=i;
new_row[i]=9999;
new_row[i+max_width]=9999;
}
old_row[2*max_width]=9999;
new_row[2*max_width]=9999;
int start_new_row = max_width-1;
for (i=1; i<=pep_len1; i++)
{
int j;
if (start_new_row>0)
{
new_row[start_new_row]=i;
new_row[--start_new_row]=9999;
}
else
new_row[0] = 9999;
for (j=1; j<= 2*max_width; j++)
{
int p2_pos = i + j - max_width;
if (p2_pos<1)
continue;
if (p2_pos>pep_len2)
break;
float v1,v2,v3,dxy=0;
int p1=pep1[i-1],p2=pep2[p2_pos-1];
if (p1 != p2)
{
dxy=1;
if ( ((p1==Ile || p1==Leu) && (p2==Ile || p2==Leu)) ||
((p1==Gln || p1==Lys) && (p2==Gln || p2==Lys)) )
{
dxy=0;
}
else if ( ((p1==Asn && p2==Asp) || (p1==Asp && p2==Asn)) ||
((pep1[i]==Gln && p2==Glu) || (p1==Glu && p2==Gln)) ||
((pep1[i]==Lys && p2==Glu) || (p1==Glu && p2==Lys)) ||
((pep1[i]==Ile && p2==Asn) || (p1==Asn && p2==Ile)) ||
((pep1[i]==Leu && p2==Asn) || (p1==Asn && p2==Leu)) )
{
dxy=0.5;
}
else if (ox_met_aa>0 && (( p1 == ox_met_aa && p2 == Phe) ||
( p1 == Phe && p2 == ox_met_aa)) )
{
dxy = 0;
}
}
v1= old_row[j]+dxy;
v2= new_row[j-1]+1;
v3= (p2_pos<pep_len2 && j<2*max_width) ? old_row[j+1]+1 : 9999;
new_row[j]=v1;
if (new_row[j]>v2)
new_row[j]=v2;
if (new_row[j]>v3)
new_row[j]=v3;
}
float *tmp;
tmp = old_row;
old_row = new_row;
new_row = tmp;
}
return old_row[pep_len2-pep_len1+max_width];
}