/// \brief TODOCUMENT
pair<alignment, superpose_orderer> ssap_scores_file_alignment_acquirer::do_get_alignment_and_orderer(const pdb_list &arg_pdbs ///< TODOCUMENT
) const {
// Parse the SSAP scores file
const path ssap_scores_file = get_ssap_scores_file();
const pair<str_vec, size_size_pair_doub_map> ssap_scores_data = ssap_scores_file::parse_ssap_scores_file( ssap_scores_file );
const str_vec &names = ssap_scores_data.first;
const size_size_pair_doub_map &scores = ssap_scores_data.second;
if ( names.size() != arg_pdbs.size() ) {
BOOST_THROW_EXCEPTION(runtime_error_exception(
"The number of PDBs is "
+ ::std::to_string( arg_pdbs.size() )
+ ", which doesn't match the "
+ ::std::to_string( names.size() )
+ " structures required for combining with the SSAP scores file \""
+ ssap_scores_file.string()
+ "\""
));
}
// Make a superpose_orderer from the scores
const superpose_orderer my_orderer = make_superpose_orderer( scores );
// Use the superpose_orderer code to get a spanning tree, ordered in descending order of score
const size_size_pair_vec spanning_tree = get_spanning_tree_ordered_by_desc_score( scores );
// // TODOCUMENT
ostringstream stderr;
const size_size_alignment_tuple_vec spanning_alignments = get_spanning_alignments(
ssap_scores_file.parent_path(),
names,
arg_pdbs,
spanning_tree,
stderr
);
// TODOCUMENT
const bool single_pdb = ( arg_pdbs.size() == 1 );
const alignment new_alignment = single_pdb ? make_single_alignment( front( arg_pdbs ).get_num_residues() )
: build_alignment_from_parts( spanning_alignments, build_protein_list_of_pdb_list( arg_pdbs ) );
if ( names.size() == 0 ) {
// Return the results
return make_pair( new_alignment, my_orderer );
}
// BOOST_LOG_TRIVIAL( warning )<< "About to attempt to build protein list using data that's been read from ssap_scores_file (with " << arg_pdbs.size() << " pdbs and " << names.size() << " names)";
const protein_list proteins_of_pdbs = build_protein_list_of_pdb_list_and_names( arg_pdbs, names );
const alignment scored_new_alignment = score_alignment_copy( residue_scorer(), new_alignment, proteins_of_pdbs );
// cerr << "Did generate alignment : \n";
// cerr << horiz_align_outputter( scored_new_alignment ) << endl;
// write_alignment_as_fasta_alignment( cerr, scored_new_alignment, build_protein_list_of_pdb_list( arg_pdbs ) );
// cerr << endl;
// Return the results
return make_pair( scored_new_alignment, my_orderer );
}
/// \brief TODOCUMENT
///
/// \relates display_colourer
display_colour_spec cath::get_colour_spec(const display_colourer &arg_colourer, ///< TODOCUMENT
const pdb_list &arg_pdbs, ///< TODOCUMENT
const str_vec &arg_names, ///< TODOCUMENT
const alignment &arg_alignment ///< TODOCUMENT
) {
const alignment::size_type num_entries = arg_alignment.num_entries();
const alignment::size_type aln_length = arg_alignment.length();
if ( aln_length <= 0 || num_entries <= 0 ) {
BOOST_THROW_EXCEPTION(invalid_argument_exception("Unable to colour the alignment_context because the alignment is empty"));
}
if ( num_entries != arg_pdbs.size() ) {
BOOST_THROW_EXCEPTION(invalid_argument_exception("Unable to colour the alignment_context because the number of entries doesn't match the number of PDBs"));
}
if ( num_entries != arg_names.size() ) {
BOOST_THROW_EXCEPTION(invalid_argument_exception("Unable to colour the alignment_context because the number of entries doesn't match the number of names"));
}
auto &&result_spec = arg_colourer.get_colour_spec( alignment_context(
arg_pdbs,
arg_names,
arg_alignment
) );
return has_score_colour_handler( arg_colourer )
? adjust_display_colour_spec_copy(
std::forward< decltype( result_spec ) >( result_spec ),
get_score_colour_handler( arg_colourer ),
arg_alignment
)
: result_spec;
}
/// \brief TODOCUMENT
pair<alignment, size_size_pair_vec> alignment_acquirer::get_alignment_and_spanning_tree(const pdb_list &arg_pdbs
) const {
// Call the concrete class's implementation of do_get_alignment_and_orderer() and grab the resulting alignment and superpose_orderer
const pair<alignment, superpose_orderer> alignment_and_orderer = do_get_alignment_and_orderer(arg_pdbs);
const alignment &new_alignment = alignment_and_orderer.first;
const superpose_orderer &my_orderer = alignment_and_orderer.second;
// Check that both are of the correct size
const size_t num_pdbs = arg_pdbs.size();
if ( new_alignment.num_entries() != num_pdbs ) {
BOOST_THROW_EXCEPTION(runtime_error_exception(
"Number of entries in alignment ("
+ lexical_cast<string>( new_alignment.num_entries() )
+ ") does not match expected number ("
+ lexical_cast<string>( num_pdbs )
+ ")"
));
}
if ( my_orderer.get_num_items() != num_pdbs ) {
BOOST_THROW_EXCEPTION(runtime_error_exception(
"Number of entries in superpose_orderer ("
+ lexical_cast<string>( my_orderer.get_num_items() )
+ ") does not match expected number ("
+ lexical_cast<string>( num_pdbs )
+ ")"
));
}
// Try to construct a spanning tree
// Catch any failures and report them sensibly
size_size_pair_vec spanning_tree;
try {
// spanning_tree = get_spanning_tree( my_orderer );
spanning_tree = get_spanning_tree_ordered_by_desc_score( my_orderer );
}
/// \todo Make the condition of this error message come from the method by which the orderer was populated
catch (const invalid_argument_exception &err) {
logger::log_and_exit(
logger::return_code::INSUFFICIENT_RESIDUE_NAME_OVERLAPS,
"Cannot construct a tree connecting all PDBs with pairs overlapping by at least " + lexical_cast<string>(alignment_acquirer::MIN_NUM_COMMON_RESIDUES_TO_SUPERPOSE_PAIR) + " residues"
);
}
// If the size of the spanning tree isn't one less than the number of PDBs then throw a wobbly
//
/// \todo Also check the size of the alignment matches?
if ( spanning_tree.size() + 1 != num_pdbs ) {
BOOST_THROW_EXCEPTION(invalid_argument_exception("The spanning tree does not correctly cover the number of PDBs"));
}
// Return the resulting alignment and spanning tree
return make_pair(new_alignment, spanning_tree);
}
/// \brief TODOCUMENT
pair<alignment, superpose_orderer> fasta_aln_file_alignment_acquirer::do_get_alignment_and_orderer(const pdb_list &arg_pdbs ///< TODOCUMENT
) const {
// Construct an alignment from the FASTA alignment file
const protein_list proteins_of_pdbs = build_protein_list_of_pdb_list( arg_pdbs );
const alignment new_alignment = read_alignment_from_fasta_file( get_fasta_alignment_file(), proteins_of_pdbs, cerr );
// Construct a superpose_orderer and set arbitrary scores to ensure that the spanning tree will connect the entries
const size_t num_pdbs = arg_pdbs.size();
superpose_orderer my_orderer( num_pdbs );
for (size_t link_ctr = 1; link_ctr < num_pdbs; ++link_ctr) {
my_orderer.set_score(link_ctr, 0, 0.0);
}
const alignment scored_new_alignment = score_alignment_copy( residue_scorer(), new_alignment, proteins_of_pdbs );
// Return the results
return make_pair( scored_new_alignment, my_orderer );
}
/// \brief TODOCUMENT
///
/// \relates display_colourer
display_colour_spec cath::get_colour_spec(const display_colourer &arg_colourer, ///< TODOCUMENT
const pdb_list &arg_pdbs, ///< TODOCUMENT
const str_vec &arg_names, ///< TODOCUMENT
const alignment &arg_alignment ///< TODOCUMENT
) {
const alignment::size_type num_entries = arg_alignment.num_entries();
const alignment::size_type aln_length = arg_alignment.length();
if ( aln_length <= 0 || num_entries <= 0 ) {
BOOST_THROW_EXCEPTION(invalid_argument_exception("Unable to colour the alignment_context because the alignment is empty"));
}
if ( num_entries != arg_pdbs.size() ) {
BOOST_THROW_EXCEPTION(invalid_argument_exception("Unable to colour the alignment_context because the number of entries doesn't match the number of PDBs"));
}
if ( num_entries != arg_names.size() ) {
BOOST_THROW_EXCEPTION(invalid_argument_exception("Unable to colour the alignment_context because the number of entries doesn't match the number of names"));
}
return arg_colourer.get_colour_spec( alignment_context(
arg_pdbs,
arg_names,
arg_alignment
) );
}
/// \brief Parse a FASTA format input into an alignment
///
/// This version of read_alignment_from_fasta() doesn't take names to find within the parsed IDS
/// so it is less safe than the other version.
///
/// \relates alignment
alignment cath::align::read_alignment_from_fasta(istream &arg_istream, ///< The istream from which to read the FASTA input for parsing
const pdb_list &arg_pdbs, ///< The PDBs that TODOCUMENT
ostream &arg_stderr ///< An ostream to which any warnings should be output (currently unused)
) {
return read_alignment_from_fasta( arg_istream, get_amino_acid_lists( arg_pdbs ), str_vec( arg_pdbs.size() ), arg_stderr );
}
/// \brief Parse a FASTA format input into an alignment
///
/// This version of read_alignment_from_fasta() doesn't take names to find within the parsed IDS
/// so it is less safe than the other version.
///
/// \relates alignment
alignment cath::align::read_alignment_from_fasta_file(const path &arg_fasta_file, ///< The file from which to read the FASTA input for parsing
const pdb_list &arg_pdbs, ///< The PDBs that TODOCUMENT
ostream &arg_stderr ///< An ostream to which any warnings should be output (currently unused)
) {
return read_alignment_from_fasta_file( arg_fasta_file, arg_pdbs, str_vec( arg_pdbs.size() ), arg_stderr );
}
/// \brief TODOCUMENT
///
/// \relates alignment
///
/// CORA file format
///
/// The header consists of the following
/// - One format line '#FM CORA_FORMAT 1.1'
/// - Any number of comment lines '#CC'
/// - Total number of proteins in the alignment
/// - All CATH domain names in the alignment
/// - Total number of alignment positions
///
/// For example:
///
/// #FM CORA_FORMAT 1.1
/// #CC
/// #CC Any number of comment lines (200 characters max per line)
/// #CC
/// #CC
/// 3
/// 6insE0 1igl00 1bqt00
/// 73
///
/// The body consists of the following:
///
/// START PROT 1 PROT 2 PROT N END
/// <------------><---------><---------><---------><---------------->
/// ddddxddddxddddxddddcxcxxcxddddcxcxxcxddddcxcxxcxxxcxddddxddddxxdd
///
/// 1 0 1 0 0 0 1 A 0 0 0 0 0 0 0 0
/// 2 0 1 0 0 0 2 Y 0 0 0 0 0 0 0 0
/// 3 0 2 1B F 0 3 R 0 0 0 0 0 0 0 0
/// 4 0 3 2B V H 4 P 0 1 G 0 0 1 0 2
/// 5 1 3 3B N H 5 S 0 2 P 0 0 1 0 6
/// 6 0 3 4B Q H 6 E 0 3 E 0 0 1 0 2
///
/// START (14 characters) :
/// - Column 1: Alignment Position (dddd)
/// - Column 2: No. of position selected for structural template (dddd)
/// - Column 3: No. of proteins aligned at this position (dddd)
///
/// PROT 1,2,3... (11 characters per protein)
/// - Column 4 (7,10 etc): Residue number in PDB file (ddddc) 4 digit number
/// - + 1 character insert code
/// - Importantly the insert code is always in the same position not within
/// - the 4 characters reserved for the pdb number (see below)
/// - Column 5 (8,11 etc): Amino Acid Code (c)
/// - Column 6 (9,12 etc): Secondary Structure Assignment (c)
///
/// END (18 characters)
/// - Last Column-3: Consensus Secondary Structure Assignment (c)
/// - Last Column-2: No. of alpha residues at this position (dddd)
/// - Last Column-1: No. of beta residues at this position (dddd)
/// - Last Column: Structural Conservation Score (dd)
alignment cath::align::read_alignment_from_cath_cora_legacy_format(istream &arg_istream, ///< TODOCUMENT
const pdb_list &arg_pdbs, ///< TODOCUMENT
ostream &arg_stderr ///< TODOCUMENT
) {
const size_t CHARS_IN_MAIN_DATA_LINE_START = 14;
const size_t CHARS_IN_MAIN_DATA_LINE_PROT = 11;
const size_t CHARS_IN_MAIN_DATA_LINE_END = 18;
if (arg_pdbs.empty()) {
BOOST_THROW_EXCEPTION(invalid_argument_exception("Cannot load a CORA legacy alignment with 0 PDB entries"));
}
arg_istream.exceptions(ios::badbit);
try {
residue_name_vec_vec residue_names_of_first_chains;
for (const pdb &arg_pdb : arg_pdbs) {
residue_names_of_first_chains.push_back( arg_pdb.get_residue_names_of_first_chain__backbone_unchecked() );
}
// Check the first line is the file format line
string line_string;
getline(arg_istream, line_string);
if (!starts_with(line_string, "#FM CORA_FORMAT ")) {
BOOST_THROW_EXCEPTION(runtime_error_exception("No CORA header file format line"));
}
// Skip any comment lines
while (getline(arg_istream, line_string) && starts_with(line_string, "#CC")) {
}
// Grab the number of proteins and ensure the alignment matches
const size_t num_proteins = lexical_cast<size_t>(line_string);
if (num_proteins != arg_pdbs.size()) {
BOOST_THROW_EXCEPTION(invalid_argument_exception("Number of PDBs in CORA file is " + lexical_cast<string>(num_proteins) + ", which does not match " + lexical_cast<string>(arg_pdbs.size())));
}
const size_t num_chars_in_main_data_line = CHARS_IN_MAIN_DATA_LINE_START + (num_proteins * CHARS_IN_MAIN_DATA_LINE_PROT) + CHARS_IN_MAIN_DATA_LINE_END;
// Grab the protein names
getline( arg_istream, line_string );
trim( line_string );
const str_vec names = split_build<str_vec>( line_string, is_space() );
if ( names.size() != num_proteins ) {
BOOST_THROW_EXCEPTION(runtime_error_exception("Splitting on space does not give " + lexical_cast<string>(num_proteins) + " entries in CORA alignment names line: \"" + line_string + "\""));
}
// Grab the total number of alignment positions
getline(arg_istream, line_string);
const size_t num_positions = lexical_cast<size_t>(line_string);
// Prepare the data structures to populate
aln_posn_vec posns( num_proteins, 0 );
opt_score_vec scores;
scores.reserve( num_positions );
opt_aln_posn_vec_vec data( num_proteins );
for (opt_aln_posn_vec &data_col : data) {
data_col.reserve( num_positions );
}
// Loop over the main data section
while (getline(arg_istream, line_string)) {
// Check the line is of the correct length
if (line_string.length() != num_chars_in_main_data_line) {
BOOST_THROW_EXCEPTION(runtime_error_exception("Number of characters in main data line does not equal " + lexical_cast<string>(num_chars_in_main_data_line)));
}
// Grab the global details from start of this line
const size_t alignment_posn = lexical_cast<size_t>( trim_copy( line_string.substr( 0, 4 ))); // Column 1: Alignment Position (dddd)
// const size_t num_entries_in_temp = lexical_cast<size_t>( trim_copy( line_string.substr( 5, 4 ))); // Column 2: No. of position selected for structural template (dddd)
const size_t num_entries_in_posn = lexical_cast<size_t>( trim_copy( line_string.substr( 10, 4 ))); // Column 2: No. of position selected for structural template (dddd)
if (alignment_posn != data.front().size() + 1) {
BOOST_THROW_EXCEPTION(runtime_error_exception("Alignment position counter " + lexical_cast<string>(alignment_posn) + " does not match " + lexical_cast<string>(data.front().size() + 1)));
}
// Loop over the indices of the proteins
size_t num_present_posns(0);
for (size_t prot_ctr = 0; prot_ctr < num_proteins; ++prot_ctr) {
// Prepare string and other data for this protein
const size_t prot_string_offset = CHARS_IN_MAIN_DATA_LINE_START + prot_ctr * CHARS_IN_MAIN_DATA_LINE_PROT;
const string prot_string = line_string.substr( prot_string_offset, CHARS_IN_MAIN_DATA_LINE_PROT );
opt_aln_posn_vec &data_col = data [ prot_ctr ];
aln_posn_type &posn = posns[ prot_ctr ];
// Grab the the details for this protein
const int residue_num = lexical_cast<int>( trim_copy( prot_string.substr( 1, 4 ))); // Column 4 (7,10 etc): Residue number in PDB file (ddddc) 4 digit number
const char insert_code = prot_string.at( 5 ) ; // + 1 character insert code
const char amino_acid = prot_string.at( 7 ) ; // Column 5 (8,11 etc): Amino Acid Code (c)
// const char sec_struc = prot_string.at( 10 ) ; // Column 6 (9,12 etc): Secondary Structure Assignment (c)
// Find the residue in the list of this PDB's residue names
const residue_name_vec &residues_names = residue_names_of_first_chains[ prot_ctr ];
const residue_name res_name = make_residue_name_with_non_insert_char( residue_num, insert_code, ' ' );
const opt_aln_posn find_result = search_for_residue_in_residue_names(
posn,
residues_names,
amino_acid,
res_name,
arg_stderr
);
data_col.push_back( find_result ? opt_aln_posn( ( *find_result ) + 1 ) : opt_aln_posn( none ) );
if ( find_result ) {
posn = *find_result;
++num_present_posns;
}
}
if (num_present_posns != num_entries_in_posn) {
BOOST_THROW_EXCEPTION(runtime_error_exception(
"Number of positions for alignment_posn " + lexical_cast<string>(alignment_posn)
+ " was " + lexical_cast<string>(num_present_posns)
+ " not " + lexical_cast<string>(num_entries_in_posn)
));
}
// Prepare the string for the global details at the end of this line
const size_t end_string_offset = CHARS_IN_MAIN_DATA_LINE_START + num_proteins * CHARS_IN_MAIN_DATA_LINE_PROT;
const string end_string = line_string.substr( end_string_offset, CHARS_IN_MAIN_DATA_LINE_END );
// Grab the global details from start of this line
// const size_t cons_sec_struc = end_string.at( 3 ) ; // Last Column-3: Consensus Secondary Structure Assignment (c)
// const size_t num_alpha_res = lexical_cast<size_t>( trim_copy( end_string.substr( 5, 4 ))); // Last Column-2: No. of alpha residues at this position (dddd)
// const size_t num_beta_res = lexical_cast<size_t>( trim_copy( end_string.substr( 10, 4 ))); // Last Column-1: No. of beta residues at this position (dddd)
const size_t cons_score = lexical_cast<size_t>( trim_copy( end_string.substr( 16, 2 ))); // Last Column: Structural Conservation Score (dd)
scores.push_back( numeric_cast<double>( cons_score ) );
// // If there are multiple entries in this position then store the score
// if (num_entries_in_posn > 1) {
//// cerr << "Adding score for " << alignment_posn-1 << endl;
// scores.push_back(cons_score);
// }
}
if ( num_positions != data.front().size() ) {
BOOST_THROW_EXCEPTION(runtime_error_exception(
"CORA legacy alignment number of positions was "
+ lexical_cast<string>( data.front().size() )
+ " not "
+ lexical_cast<string>( num_positions )
) );
}
alignment new_alignment = alignment_offset_1_factory( data );
// Create a scores matrix and then empty any cells that are absent from the alignment
opt_score_vec_vec all_scores( new_alignment.num_entries(), scores );
for (size_t entry = 0; entry < new_alignment.num_entries(); ++entry) {
for (size_t index = 0; index < new_alignment.length(); ++index) {
if ( ! has_position_of_entry_of_index( new_alignment, entry, index ) ) {
all_scores[ entry ][ index ] = none;
}
}
}
set_scores( new_alignment, all_scores);
return new_alignment;
}
// Catch any I/O exceptions
catch (const std::exception &ex) {
const string error_message(string("Cannot read CORA legacy alignment file [") + ex.what() + "] ");
perror(error_message.c_str());
BOOST_THROW_EXCEPTION(runtime_error_exception(error_message));
};
}