HwRMSD.cpp

#include "HwRMSD.h"

using namespace std;

void print_extra_help()
{
    cout <<
"Additional options:\n"
"    -dir     Perform all-against-all alignment among the list of PDB\n"
"             chains listed by 'chain_list' under 'chain_folder'. Note\n"
"             that the slash is necessary.\n"
"             $ HwRMSD -dir chain_folder/ chain_list\n"
"\n"
"    -dir1    Use chain2 to search a list of PDB chains listed by 'chain1_list'\n"
"             under 'chain1_folder'. Note that the slash is necessary.\n"
"             $ HwRMSD -dir1 chain1_folder/ chain1_list chain2\n"
"\n"
"    -dir2    Use chain1 to search a list of PDB chains listed by 'chain2_list'\n"
"             under 'chain2_folder'\n"
"             $ HwRMSD chain1 -dir2 chain2_folder/ chain2_list\n"
"\n"
"    -suffix  (Only when -dir1 and/or -dir2 are set, default is empty)\n"
"             add file name suffix to files listed by chain1_list or chain2_list\n"
"\n"
"    -atom    4-character atom name used to represent a residue.\n"
"             Default is \" C3'\" for RNA/DNA and \" CA \" for proteins\n"
"             (note the spaces before and after CA).\n"
"\n"
"    -mol     Molecule type: RNA or protein\n"
"             Default is detect molecule type automatically\n"
"\n"
"    -ter     Strings to mark the end of a chain\n"
"             3: (default) TER, ENDMDL, END or different chain ID\n"
"             2: ENDMDL, END, or different chain ID\n"
"             1: ENDMDL or END\n"
"             0: (default in the first C++ TMalign) end of file\n"
"\n"
"    -split   Whether to split PDB file into multiple chains\n"
"             0: (default) treat the whole structure as one single chain\n"
"             1: treat each MODEL as a separate chain (-ter should be 0)\n"
"             2: treat each chain as a seperate chain (-ter should be <=1)\n"
"\n"
"    -outfmt  Output format\n"
"             0: (default) full output\n"
"             1: fasta format compact output\n"
"             2: tabular format very compact output\n"
"            -1: full output, but without version or citation information\n"
"\n"
"    -byresi  Whether to assume residue index correspondence between the\n"
"             two structures.\n"
"             0: (default) sequence independent alignment\n"
"             1: (same as TMscore program) sequence-dependent superposition,\n"
"                i.e. align by residue index\n"
"             2: (same as TMscore -c, should be used with -ter <=1)\n"
"                align by residue index and chain ID\n"
"             3: (similar to TMscore -c, should be used with -ter <=1)\n"
"                align by residue index and order of chain\n"
"\n"
"    -glocal  Global or local alignment\n"
"             0: (default) Needleman-Wunsch algorithm for global alignment\n"
"             1: glocal-query alignment\n"
"             2: glocal-both alignment\n"
"             3: Smith-Waterman algorithm for local alignment\n"
"\n"
"    -iter    Alignment-superposition iterations. Default is 10\n"
"\n"
"    -seq     Type of sequence used to make initial alignment\n"
"             1: amino acid/nucleotide sequence\n"
"             2: secondary structure\n"
"             3: (default) sequence + secondary structure\n"
"\n"
"    -het     Whether to align residues marked as 'HETATM' in addition to 'ATOM  '\n"
"             0: (default) only align 'ATOM  ' residues\n"
"             1: align both 'ATOM  ' and 'HETATM' residues\n"
"\n"
"    -infmt1  Input format for chain1\n"
"    -infmt2  Input format for chain2\n"
"            -1: (default) automatically detect PDB or PDBx/mmCIF format\n"
"             0: PDB format\n"
"             1: SPICKER format\n"
"             2: xyz format\n"
"             3: PDBx/mmCIF format\n"
    <<endl;
}

void print_help(bool h_opt=false)
{
    cout <<
"Partial implement of HwRMSD method for sequence-guided structure alignment.\n"
"\n"
"Usage: HwRMSD PDB1.pdb PDB2.pdb [Options]\n"
"\n"
"Options:\n"
"    -u    TM-score normalized by user assigned length (the same as -L)\n"
"          warning: it should be >= minimum length of the two structures\n"
"          otherwise, TM-score may be >1\n"
"\n"
"    -a    TM-score normalized by the average length of two structures\n"
"          T or F, (default F)\n"
"\n"
"    -i    Start with an alignment specified in fasta file 'align.txt'\n"
"\n"
"    -I    Stick to the alignment specified in 'align.txt'\n"
"\n"
"    -m    Output HwRMSD rotation matrix\n"
"\n"
"    -d    TM-score scaled by an assigned d0, e.g. 5 Angstroms\n"
"\n"
"    -o    Output the superposition of PDB1.pdb to TM_sup.pdb\n"
"          $ HwRMSD PDB1.pdb PDB2.pdb -o TM_sup.pdb\n"
"          To view superposed full-atom structures:\n"
"          $ pymol TM_sup.pdb PDB2.pdb\n"
"\n"
"    -h    Print the full help message\n"
"\n"
"    (Options -u, -a, -d, -o won't change the final structure alignment)\n\n"
"Example usages:\n"
"    HwRMSD PDB1.pdb PDB2.pdb\n"
"    HwRMSD PDB1.pdb PDB2.pdb -u 100 -d 5.0\n"
"    HwRMSD PDB1.pdb PDB2.pdb -a T -o PDB1.sup\n"
"    HwRMSD PDB1.pdb PDB2.pdb -i align.txt\n"
"    HwRMSD PDB1.pdb PDB2.pdb -m matrix.txt\n"
    <<endl;

    if (h_opt) print_extra_help();

    exit(EXIT_SUCCESS);
}

int main(int argc, char *argv[])
{
    if (argc < 2) print_help();

    /**********************/
    /*    get argument    */
    /**********************/
    string xname       = "";
    string yname       = "";
    string fname_super = ""; // file name for superposed structure
    string fname_lign  = ""; // file name for user alignment
    string fname_matrix= ""; // file name for output matrix
    vector<string> sequence; // get value from alignment file
    double Lnorm_ass, d0_scale;

    bool h_opt = false; // print full help message
    bool m_opt = false; // flag for -m, output rotation matrix
    int  i_opt = 0;     // 0 for -i, 3 for -I
    bool o_opt = false; // flag for -o, output superposed structure
    bool a_opt = false; // flag for -a, normalized by average length
    bool u_opt = false; // flag for -u, normalized by user specified length
    bool d_opt = false; // flag for -d, user specified d0

    int    infmt1_opt=-1;    // PDB or PDBx/mmCIF format for chain_1
    int    infmt2_opt=-1;    // PDB or PDBx/mmCIF format for chain_2
    int    ter_opt   =3;     // TER, END, or different chainID
    int    split_opt =0;     // do not split chain
    int    outfmt_opt=0;     // set -outfmt to full output
    int    het_opt=0;        // do not read HETATM residues
    string atom_opt  ="auto";// use C alpha atom for protein and C3' for RNA
    string mol_opt   ="auto";// auto-detect the molecule type as protein/RNA
    string suffix_opt="";    // set -suffix to empty
    string dir_opt   ="";    // set -dir to empty
    string dir1_opt  ="";    // set -dir1 to empty
    string dir2_opt  ="";    // set -dir2 to empty
    int    byresi_opt=0;     // set -byresi to 0
    vector<string> chain1_list; // only when -dir1 is set
    vector<string> chain2_list; // only when -dir2 is set
    vector<string> chain2parse1;
    vector<string> chain2parse2;
    vector<string> model2parse1;
    vector<string> model2parse2;
    int    glocal    =0;
    int    iter_opt  =10;
    double early_opt =0.01;
    int    seq_opt   =3;

    for(int i = 1; i < argc; i++)
    {
        if ( !strcmp(argv[i],"-o") && i < (argc-1) )
        {
            fname_super = argv[i + 1];     o_opt = true; i++;
        }
        else if ( (!strcmp(argv[i],"-u") || 
                   !strcmp(argv[i],"-L")) && i < (argc-1) )
        {
            Lnorm_ass = atof(argv[i + 1]); u_opt = true; i++;
        }
        else if ( !strcmp(argv[i],"-a") && i < (argc-1) )
        {
            if (!strcmp(argv[i + 1], "T"))      a_opt=true;
            else if (!strcmp(argv[i + 1], "F")) a_opt=false;
            else PrintErrorAndQuit("Wrong value for option -a! It should be T or F");
            i++;
        }
        else if ( !strcmp(argv[i],"-d") && i < (argc-1) )
        {
            d0_scale = atof(argv[i + 1]); d_opt = true; i++;
        }
        else if ( !strcmp(argv[i],"-h") )
        {
            h_opt = true;
        }
        else if ( !strcmp(argv[i],"-i") && i < (argc-1) )
        {
            if (i_opt==3)
                PrintErrorAndQuit("ERROR! -i and -I cannot be used together");
            fname_lign = argv[i + 1];      i_opt = 1; i++;
        }
        else if (!strcmp(argv[i], "-I") && i < (argc-1) )
        {
            if (i_opt==1)
                PrintErrorAndQuit("ERROR! -I and -i cannot be used together");
            fname_lign = argv[i + 1];      i_opt = 3; i++;
        }
        else if (!strcmp(argv[i], "-m") && i < (argc-1) )
        {
            fname_matrix = argv[i + 1];    m_opt = true; i++;
        }// get filename for rotation matrix
        else if ( !strcmp(argv[i],"-infmt1") && i < (argc-1) )
        {
            infmt1_opt=atoi(argv[i + 1]); i++;
        }
        else if ( !strcmp(argv[i],"-infmt2") && i < (argc-1) )
        {
            infmt2_opt=atoi(argv[i + 1]); i++;
        }
        else if ( !strcmp(argv[i],"-ter") && i < (argc-1) )
        {
            ter_opt=atoi(argv[i + 1]); i++;
        }
        else if ( !strcmp(argv[i],"-split") && i < (argc-1) )
        {
            split_opt=atoi(argv[i + 1]); i++;
        }
        else if ( !strcmp(argv[i],"-atom") && i < (argc-1) )
        {
            atom_opt=argv[i + 1]; i++;
        }
        else if ( !strcmp(argv[i],"-mol") && i < (argc-1) )
        {
            mol_opt=argv[i + 1]; i++;
        }
        else if ( !strcmp(argv[i],"-dir") && i < (argc-1) )
        {
            dir_opt=argv[i + 1]; i++;
        }
        else if ( !strcmp(argv[i],"-dir1") && i < (argc-1) )
        {
            dir1_opt=argv[i + 1]; i++;
        }
        else if ( !strcmp(argv[i],"-dir2") && i < (argc-1) )
        {
            dir2_opt=argv[i + 1]; i++;
        }
        else if ( !strcmp(argv[i],"-suffix") && i < (argc-1) )
        {
            suffix_opt=argv[i + 1]; i++;
        }
        else if ( !strcmp(argv[i],"-outfmt") && i < (argc-1) )
        {
            outfmt_opt=atoi(argv[i + 1]); i++;
        }
        else if ( !strcmp(argv[i],"-byresi") && i < (argc-1) )
        {
            byresi_opt=atoi(argv[i + 1]); i++;
        }
        else if ( !strcmp(argv[i],"-glocal") && i < (argc-1) )
        {
            glocal=atoi(argv[i + 1]); i++;
        }
        else if ( !strcmp(argv[i],"-iter") && i < (argc-1) )
        {
            iter_opt=atoi(argv[i + 1]); i++;
        }
        else if ( !strcmp(argv[i],"-early") && i < (argc-1) )
        {
            early_opt=atof(argv[i + 1]); i++;
        }
        else if ( !strcmp(argv[i],"-seq") && i < (argc-1) )
        {
            seq_opt=atoi(argv[i + 1]); i++;
        }
        else if ( !strcmp(argv[i],"-het") && i < (argc-1) )
        {
            het_opt=atoi(argv[i + 1]); i++;
        }
        else if (!strcmp(argv[i], "-chain1") )
        {
            if (i>=(argc-1)) 
                PrintErrorAndQuit("ERROR! Missing value for -chain1");
            split(argv[i+1],chain2parse1,',');
            i++;
        }
        else if (!strcmp(argv[i], "-chain2") )
        {
            if (i>=(argc-1)) 
                PrintErrorAndQuit("ERROR! Missing value for -chain2");
            split(argv[i+1],chain2parse2,',');
            i++;
        }
        else if (!strcmp(argv[i], "-model1") )
        {
            if (i>=(argc-1)) 
                PrintErrorAndQuit("ERROR! Missing value for -model1");
            split(argv[i+1],model2parse1,',');
            i++;
        }
        else if (!strcmp(argv[i], "-model2") )
        {
            if (i>=(argc-1)) 
                PrintErrorAndQuit("ERROR! Missing value for -model2");
            split(argv[i+1],model2parse2,',');
            i++;
        }
        else if (xname.size() == 0) xname=argv[i];
        else if (yname.size() == 0) yname=argv[i];
        else PrintErrorAndQuit(string("ERROR! Undefined option ")+argv[i]);
    }

    if(xname.size()==0 || (yname.size()==0 && dir_opt.size()==0) || 
                          (yname.size()    && dir_opt.size()))
    {
        if (h_opt) print_help(h_opt);
        if (xname.size()==0)
            PrintErrorAndQuit("Please provide input structures");
        else if (yname.size()==0 && dir_opt.size()==0)
            PrintErrorAndQuit("Please provide structure B");
        else if (yname.size() && dir_opt.size())
            PrintErrorAndQuit("Please provide only one file name if -dir is set");
    }

    if (suffix_opt.size() && dir_opt.size()+dir1_opt.size()+dir2_opt.size()==0)
        PrintErrorAndQuit("-suffix is only valid if -dir, -dir1 or -dir2 is set");
    if ((dir_opt.size() || dir1_opt.size() || dir2_opt.size()))
    {
        if (m_opt || o_opt)
            PrintErrorAndQuit("-m or -o cannot be set with -dir, -dir1 or -dir2");
        else if (dir_opt.size() && (dir1_opt.size() || dir2_opt.size()))
            PrintErrorAndQuit("-dir cannot be set with -dir1 or -dir2");
    }

    bool autojustify=(atom_opt=="auto" || atom_opt=="PC4'"); // auto re-pad atom name
    if (atom_opt.size()!=4)
        PrintErrorAndQuit("ERROR! Atom name must have 4 characters, including space.");
    if (mol_opt!="auto" && mol_opt!="protein" && mol_opt!="RNA")
        PrintErrorAndQuit("ERROR! Molecule type must be either RNA or protein.");
    else if (mol_opt=="protein" && atom_opt=="auto")
        atom_opt=" CA ";
    else if (mol_opt=="RNA" && atom_opt=="auto")
        atom_opt=" C3'";

    if (u_opt && Lnorm_ass<=0)
        PrintErrorAndQuit("Wrong value for option -u!  It should be >0");
    if (d_opt && d0_scale<=0)
        PrintErrorAndQuit("Wrong value for option -d!  It should be >0");
    if (outfmt_opt>=2 && (a_opt || u_opt || d_opt))
        PrintErrorAndQuit("-outfmt 2 cannot be used with -a, -u, -L, -d");
    if (byresi_opt!=0)
    {
        if (i_opt)
            PrintErrorAndQuit("-byresi >=1 cannot be used with -i or -I");
        if (byresi_opt<0 || byresi_opt>3)
            PrintErrorAndQuit("-byresi can only be 0, 1, 2 or 3");
        if (byresi_opt>=2 && ter_opt>=2)
            PrintErrorAndQuit("-byresi >=2 should be used with -ter <=1");
    }
    if (split_opt==1 && ter_opt!=0)
        PrintErrorAndQuit("-split 1 should be used with -ter 0");
    else if (split_opt==2 && ter_opt!=0 && ter_opt!=1)
        PrintErrorAndQuit("-split 2 should be used with -ter 0 or 1");
    if (split_opt<0 || split_opt>2)
        PrintErrorAndQuit("-split can only be 0, 1 or 2");
    if (iter_opt<=0) PrintErrorAndQuit("-iter must be >0");
    if (seq_opt!=1 && seq_opt!=2 && seq_opt!=3)
        PrintErrorAndQuit("-seq must be 1, 2 or 3");

    /* read initial alignment file from 'align.txt' */
    if (i_opt) read_user_alignment(sequence, fname_lign, i_opt);

    if (byresi_opt) i_opt=3;

    if (m_opt && fname_matrix == "") // Output rotation matrix: matrix.txt
        PrintErrorAndQuit("ERROR! Please provide a file name for option -m!");

    /* parse file list */
    if (dir1_opt.size()+dir_opt.size()==0) chain1_list.push_back(xname);
    else file2chainlist(chain1_list, xname, dir_opt+dir1_opt, suffix_opt);

    if (dir_opt.size())
        for (int i=0;i<chain1_list.size();i++)
            chain2_list.push_back(chain1_list[i]);
    else if (dir2_opt.size()==0) chain2_list.push_back(yname);
    else file2chainlist(chain2_list, yname, dir2_opt, suffix_opt);

    if (outfmt_opt==2)
        cout<<"#PDBchain1\tPDBchain2\tTM1\tTM2\t"
            <<"RMSD\tID1\tID2\tIDali\tL1\tL2\tLali"<<endl;

    /* declare previously global variables */
    vector<vector<string> >PDB_lines1; // text of chain1
    vector<vector<string> >PDB_lines2; // text of chain2
    vector<int> mol_vec1;              // molecule type of chain1, RNA if >0
    vector<int> mol_vec2;              // molecule type of chain2, RNA if >0
    vector<string> chainID_list1;      // list of chainID1
    vector<string> chainID_list2;      // list of chainID2
    int    i,j;                // file index
    int    chain_i,chain_j;    // chain index
    int    xlen, ylen;         // chain length
    int    xchainnum,ychainnum;// number of chains in a PDB file
    char   *seqx, *seqy;       // for the protein sequence 
    char    *secx, *secy;       // for the secondary structure 
    double **xa, **ya;         // for input vectors xa[0...xlen-1][0..2] and
                               // ya[0...ylen-1][0..2], in general,
                               // ya is regarded as native structure 
                               // --> superpose xa onto ya
    vector<string> resi_vec1;  // residue index for chain1
    vector<string> resi_vec2;  // residue index for chain2

    /* loop over file names */
    for (int i=0;i<chain1_list.size();i++)
    {
        /* parse chain 1 */
        xname=chain1_list[i];
        xchainnum=get_PDB_lines(xname, PDB_lines1, chainID_list1, mol_vec1,
            ter_opt, infmt1_opt, atom_opt, autojustify, split_opt, het_opt,
            chain2parse1,model2parse1);
        if (!xchainnum)
        {
            cerr<<"Warning! Cannot parse file: "<<xname
                <<". Chain number 0."<<endl;
            continue;
        }
        for (int chain_i=0;chain_i<xchainnum;chain_i++)
        {
            xlen=PDB_lines1[chain_i].size();
            if (mol_opt=="RNA") mol_vec1[chain_i]=1;
            else if (mol_opt=="protein") mol_vec1[chain_i]=-1;
            if (!xlen)
            {
                cerr<<"Warning! Cannot parse file: "<<xname
                    <<". Chain length 0."<<endl;
                continue;
            }
            NewArray(&xa, xlen, 3);
            seqx = new char[xlen + 1];
            xlen = read_PDB(PDB_lines1[chain_i], xa, seqx, 
                resi_vec1, byresi_opt);
            if (seq_opt==2 || (seq_opt==3 && iter_opt>=2))  // SS assignment
            {
                secx = new char[xlen+1];
                if (mol_vec1[chain_i]>0) 
                     make_sec(seqx, xa, xlen, secx,atom_opt);
                else make_sec(xa, xlen, secx);
            }

            for (int j=(dir_opt.size()>0)*(i+1);j<chain2_list.size();j++)
            {
                /* parse chain 2 */
                if (PDB_lines2.size()==0)
                {
                    yname=chain2_list[j];
                    ychainnum=get_PDB_lines(yname, PDB_lines2, chainID_list2,
                        mol_vec2, ter_opt, infmt2_opt, atom_opt, autojustify,
                        split_opt, het_opt, chain2parse2,model2parse2);
                    if (!ychainnum)
                    {
                        cerr<<"Warning! Cannot parse file: "<<yname
                            <<". Chain number 0."<<endl;
                        continue;
                    }
                }
                for (int chain_j=0;chain_j<ychainnum;chain_j++)
                {
                    ylen=PDB_lines2[chain_j].size();
                    if (mol_opt=="RNA") mol_vec2[chain_j]=1;
                    else if (mol_opt=="protein") mol_vec2[chain_j]=-1;
                    if (!ylen)
                    {
                        cerr<<"Warning! Cannot parse file: "<<yname
                            <<". Chain length 0."<<endl;
                        continue;
                    }
                    NewArray(&ya, ylen, 3);
                    seqy = new char[ylen + 1];
                    ylen = read_PDB(PDB_lines2[chain_j], ya, seqy,
                        resi_vec2, byresi_opt);
                    if (seq_opt==2 || (seq_opt==3 && iter_opt>=2))  // SS assignment
                    {
                        secy = new char[ylen+1];
                        if (mol_vec2[chain_j]>0)
                             make_sec(seqy, ya, ylen, secy, atom_opt);
                        else make_sec(ya, ylen, secy);
                    }

                    if (byresi_opt) extract_aln_from_resi(sequence,
                        seqx,seqy,resi_vec1,resi_vec2,byresi_opt);

                    /* declare variable specific to this pair of HwRMSD */
                    double t0[3], u0[3][3];
                    double TM1, TM2;
                    double TM3, TM4, TM5;     // for a_opt, u_opt, d_opt
                    double d0_0, TM_0;
                    double d0A, d0B, d0u, d0a;
                    double d0_out=5.0;
                    string seqM, seqxA, seqyA;// for output alignment
                    double rmsd0 = 0.0;
                    int L_ali;                // Aligned length in standard_TMscore
                    double Liden=0;
                    double TM_ali, rmsd_ali;  // TMscore and rmsd in standard_TMscore
                    int n_ali=0;
                    int n_ali8=0;
                    int *invmap = new int[ylen+1];

                    /* entry function for structure alignment */
                    HwRMSD_main(xa, ya, seqx, seqy, secx, secy, t0, u0,
                        TM1, TM2, TM3, TM4, TM5, d0_0, TM_0,
                        d0A, d0B, d0u, d0a, d0_out, seqM, seqxA, seqyA,
                        rmsd0, L_ali, Liden, TM_ali,
                        rmsd_ali, n_ali, n_ali8, xlen, ylen, sequence,
                        Lnorm_ass, d0_scale, i_opt, a_opt, u_opt, d_opt,
                        mol_vec1[chain_i]+mol_vec2[chain_j],
                        invmap, glocal, iter_opt, seq_opt, early_opt);

                    if (outfmt_opt>=2) 
                        get_seqID(invmap, seqx, seqy, ylen, Liden, n_ali8);

                    /* print result */
                    output_results(
                        xname.substr(dir1_opt.size()+dir_opt.size()),
                        yname.substr(dir2_opt.size()+dir_opt.size()),
                        chainID_list1[chain_i].c_str(),
                        chainID_list2[chain_j].c_str(),
                        xlen, ylen, t0, u0, TM1, TM2, 
                        TM3, TM4, TM5, rmsd0, d0_out,
                        seqM.c_str(), seqxA.c_str(), seqyA.c_str(), Liden,
                        n_ali8, L_ali, TM_ali, rmsd_ali,
                        TM_0, d0_0, d0A, d0B,
                        Lnorm_ass, d0_scale, d0a, d0u, 
                        (m_opt?fname_matrix+chainID_list1[chain_i]:"").c_str(),
                        outfmt_opt, ter_opt, false, split_opt, o_opt,
                        (o_opt?fname_super+chainID_list1[chain_i]:"").c_str(),
                        false, a_opt, u_opt, d_opt, 0,
                        resi_vec1, resi_vec2);

                    /* Done! Free memory */
                    seqM.clear();
                    seqxA.clear();
                    seqyA.clear();
                    DeleteArray(&ya, ylen);
                    delete [] seqy;
                    delete [] invmap;
                    if (iter_opt>=2) delete [] secy;
                    resi_vec2.clear();
                } // chain_j
                if (chain2_list.size()>1)
                {
                    yname.clear();
                    for (int chain_j=0;chain_j<ychainnum;chain_j++)
                        PDB_lines2[chain_j].clear();
                    PDB_lines2.clear();
                    chainID_list2.clear();
                    mol_vec2.clear();
                }
            } // j
            PDB_lines1[chain_i].clear();
            DeleteArray(&xa, xlen);
            delete [] seqx;
            if (iter_opt>=2) delete [] secx;
            resi_vec1.clear();
        } // chain_i
        xname.clear();
        PDB_lines1.clear();
        chainID_list1.clear();
        mol_vec1.clear();
    } // i
    if (chain2_list.size()==1)
    {
        yname.clear();
        for (int chain_j=0;chain_j<ychainnum;chain_j++)
            PDB_lines2[chain_j].clear();
        PDB_lines2.clear();
        resi_vec2.clear();
        chainID_list2.clear();
        mol_vec2.clear();
    }
    chain1_list.clear();
    chain2_list.clear();
    vector<string>().swap(chain2parse1);
    vector<string>().swap(chain2parse2);
    vector<string>().swap(model2parse1);
    vector<string>().swap(model2parse2);
    sequence.clear();
    return 0;
}