/// \brief Construct the permutation represented by changing from the original amino acid ordering to the new amino acid ordering
///
/// \todo Tidied up this code
size_vec substitution_matrix::order_permutation(const amino_acid_vec &arg_orig_aa_ordering, ///< The original amino acid ordering
const amino_acid_vec &arg_new_aa_ordering ///< The new amino acid ordering
) {
// Grab the number of amino acids and check it's consistent
const size_t num_amino_acids = arg_orig_aa_ordering.size();
if ( num_amino_acids != arg_new_aa_ordering.size() ) {
BOOST_THROW_EXCEPTION(invalid_argument_exception("Mismatch in the number of amino acids"));
}
// Construct a permutation and reserve the memory
size_vec permutation;
permutation.reserve( num_amino_acids );
// For each of the amino acids in the original ordering...
for (const amino_acid &orig_amino_acid : arg_orig_aa_ordering) {
// Find the index of the amino acid in the new ordering
const size_t index_in_new = numeric_cast<size_t>( distance(
common::cbegin( arg_new_aa_ordering ),
find(
arg_new_aa_ordering,
orig_amino_acid
)
) );
// ...check the index is valid...
if ( index_in_new >= num_amino_acids ) {
BOOST_THROW_EXCEPTION(invalid_argument_exception("Amino acid not found in new amino acids"));
}
// ...and then add the index to the back of the permutation
permutation.push_back( index_in_new );
}
// Return the resulting permutation
return permutation;
}
/// \brief Align a sequence against a corresponding pdb
/// (broadly handling residues missing in the sequence but not extra residues)
///
/// \returns A vector of opt_aln_posns corresponding to the letters of the sequence. Each is:
/// * none if the entry is a '-' character
/// * the index of the corresponding residue in arg_pdb otherwise
opt_aln_posn_vec cath::align::align_sequence_to_amino_acids(const string &arg_sequence_string, ///< The raw sequence string (no headers; no whitespace) to be aligned
const amino_acid_vec &arg_amino_acids, ///< The PDB against which the sequence is to be aligned
const string &arg_name, ///< The name of the entry to use in warnings / errors
ostream &/*arg_stderr*/ ///< The ostream to which warnings should be output
) {
const size_t sequence_length = arg_sequence_string.length();
const size_t num_pdb_residues = arg_amino_acids.size();
// Prepare the variables to be populated when looping through the sequence
str_vec skipped_residues;
opt_aln_posn_vec new_posns;
new_posns.reserve( sequence_length );
size_t pdb_ctr = 0;
// Loop along the sequence
for (size_t seq_ctr = 0; seq_ctr < sequence_length; ++seq_ctr) {
const char &sequence_char = arg_sequence_string[ seq_ctr ];
// If this is a '-' character then add none to the back of new_posns
if ( sequence_char == '-' ) {
new_posns.push_back( none );
}
// Otherwise, it's an amino-acid letter
else {
// Continue searching pdb_ctr through the PDB until it matches this letter
while ( sequence_char != arg_amino_acids[ pdb_ctr ].get_letter() ) {
skipped_residues.push_back( lexical_cast<string>( pdb_ctr ) );
// Increment pdb_ctr
++pdb_ctr;
// If pdb_ctr has overrun the end, then throw an exception
if ( pdb_ctr >= num_pdb_residues ) {
BOOST_THROW_EXCEPTION(runtime_error_exception(
"When aligning a sequence to a PDB for "
+ arg_name
+ ", could not find match in PDB for residue '"
+ sequence_char
+ "' at position "
+ lexical_cast<string>( seq_ctr )
));
}
}
// Add the found PDB position to the back of new_positions
new_posns.push_back( pdb_ctr );
// Increment the pdb_ctr to the next residue
++pdb_ctr;
}
}
const size_t num_posns_skipped = skipped_residues.size();
const size_t num_posns_found = num_pdb_residues - num_posns_skipped;
if ( num_posns_skipped > num_pdb_residues ) {
BOOST_THROW_EXCEPTION(runtime_error_exception("The number of residues skipped exceeds the total number of residues"));
}
// If the number of residues found is an unacceptably low fraction of residues in the PDB,
// then throw an exception
const double fraction_pdb_residues_found = numeric_cast<double>( num_posns_found ) / numeric_cast<double>( num_pdb_residues );
if ( fraction_pdb_residues_found < MIN_FRAC_OF_PDB_RESIDUES_IN_SEQ ) {
BOOST_THROW_EXCEPTION(runtime_error_exception(
"When aligning a sequence to a PDB for "
+ arg_name
+ ", only found matches for "
+ lexical_cast<string>( num_posns_found )
+ " of the "
+ lexical_cast<string>( pdb_ctr )
+ " residues in the PDB"
));
}
// If not all residues were found, then output a warning about
if ( num_posns_found < num_pdb_residues ) {
BOOST_LOG_TRIVIAL( warning ) << "When aligning a sequence to a PDB for \""
<< arg_name
<< "\", "
<< lexical_cast<string>( num_pdb_residues - num_posns_found )
<< " of the PDB's "
<< lexical_cast<string>( num_pdb_residues )
<< " residues were missing in the sequence and had to be inserted (residue indices, using offset of 0 : "
<< join( skipped_residues, ", " )
<< ")";
}
// Return the result of this work
return new_posns;
}