density_tools.cc

Go to the documentation of this file.

// -*- mode:c++;tab-width:2;indent-tabs-mode:t;show-trailing-whitespace:t;rm-trailing-spaces:t -*-
// vi: set ts=2 noet:
//
// (c) Copyright Rosetta Commons Member Institutions.
// (c) This file is part of the Rosetta software suite and is made available under license.
// (c) The Rosetta software is developed by the contributing members of the Rosetta Commons.
// (c) For more information, see http://www.rosettacommons.org. Questions about this can be
// (c) addressed to University of Washington UW TechTransfer, email: license@u.washington.edu.

/// @file
/// @brief

#include <protocols/viewer/viewers.hh>

#include <protocols/moves/MonteCarlo.hh>
#include <protocols/moves/Mover.hh>

#include <core/types.hh>

#include <core/chemical/AA.hh>
#include <core/chemical/ChemicalManager.hh>
#include <core/chemical/ResidueTypeSet.fwd.hh>
#include <core/scoring/ScoreFunction.hh>
#include <core/scoring/ScoreFunctionFactory.hh>
#include <core/pose/Pose.hh>
#include <core/pose/PDBInfo.hh>
#include <core/scoring/electron_density/ElectronDensity.hh>
#include <core/import_pose/import_pose.hh>
#include <protocols/electron_density/util.hh>
#include <protocols/electron_density/SetupForDensityScoringMover.hh>

#include <basic/options/util.hh>
#include <devel/init.hh>


#include <utility/vector1.hh>

#include <numeric/xyzVector.hh>
#include <numeric/random/random.hh>
#include <numeric/constants.hh>

#include <ObjexxFCL/string.functions.hh>
#include <core/kinematics/Jump.hh>

#include <basic/options/option.hh>
#include <basic/options/option_macros.hh>
#include <basic/options/keys/OptionKeys.hh>
#include <basic/options/keys/in.OptionKeys.gen.hh>
#include <basic/options/keys/edensity.OptionKeys.gen.hh>

// C++ headers
#include <iostream>
#include <fstream>
#include <string>


using namespace core;
using namespace basic;
using namespace utility;
using namespace protocols;


OPT_1GRP_KEY(File, edensity, alt_mapfile)
OPT_1GRP_KEY(Integer, denstools, nresbins)
OPT_1GRP_KEY(Integer, denstools, rscc_wdw)
OPT_1GRP_KEY(Real, denstools, lowres)
OPT_1GRP_KEY(Real, denstools, hires)
OPT_1GRP_KEY(Real, denstools, truncate_lowres)
OPT_1GRP_KEY(Real, denstools, truncate_hires)
OPT_1GRP_KEY(Boolean, denstools, truncate_map)
OPT_1GRP_KEY(Boolean, denstools, verbose)
OPT_1GRP_KEY(Boolean, denstools, super_verbose)
OPT_1GRP_KEY(Boolean, denstools, dump_map_and_mask)
OPT_1GRP_KEY(Boolean, denstools, mask)
OPT_1GRP_KEY(Boolean, denstools, mask_modelmap)
OPT_1GRP_KEY(Real, denstools, mask_radius)
OPT_1GRP_KEY(Boolean, denstools, perres)
OPT_1GRP_KEY(Boolean, denstools, bin_squared)


using core::scoring::electron_density::poseCoords;
using core::scoring::electron_density::poseCoord;

// map atom names to elements
//   loosely based on openbabel logic
std::string
name2elt( std::string line ) {
  std::string atmid = line.substr(12,4);
  while ( !atmid.empty() && atmid[0] == ' ' ) atmid = atmid.substr(1,atmid.size()-1);
  while ( !atmid.empty() && atmid[atmid.size()-1] == ' ' ) atmid = atmid.substr(0,atmid.size()-1);
  std::string resname = line.substr(17,3);
  while ( !resname.empty() && resname[0] == ' ' ) resname = resname.substr(1,resname.size()-1);
  while ( !resname.empty() && resname[resname.size()-1] == ' ' ) resname = resname.substr(0,resname.size()-1);

  std::string type;

  if ( line.substr(0,4) == "ATOM" ) {
    type = atmid.substr(0,2);
    if ( isdigit(type[0]) ) {
      // sometimes non-standard files have, e.g 11HH
      if ( !isdigit(type[1]) ) type = atmid.substr(1,1);
      else type = atmid.substr(2,1);
    } else if ( (line[12] == ' ' && type!="Zn" && type!="Fe" && type!="ZN" && type!="FE")
        || isdigit(type[1]) ) {
      type = atmid.substr(0,1);     // one-character element
    }

    if ( resname.substr(0,2) == "AS" || resname[0] == 'N' ) {
      if ( atmid == "AD1" ) type = "O";
      if ( atmid == "AD2" ) type = "N";
    }
    if ( resname.substr(0,3) == "HIS" || resname[0] == 'H' ) {
      if ( atmid == "AD1" || atmid == "AE2" ) type = "N";
      if ( atmid == "AE1" || atmid == "AD2" ) type = "C";
    }
    if ( resname.substr(0,2) == "GL" || resname[0] == 'Q' ) {
      if ( atmid == "AE1" ) type = "O";
      if ( atmid == "AE2" ) type = "N";
    }
    if ( atmid.substr(0,2) == "HH" ) { // ARG
      type = "H";
    }
    if ( atmid.substr(0,2) == "HD" || atmid.substr(0,2) == "HE" || atmid.substr(0,2) == "HG" ) {
      type = "H";
    }
  } else {
    if ( isalpha(atmid[0]) ) {
      if ( atmid.size() > 2 && (atmid[2] == '\0' || atmid[2] == ' ') ) {
        type = atmid.substr(0,2);
      } else if ( atmid[0] == 'A' ) { // alpha prefix
        type = atmid.substr(1, atmid.size() - 1);
      } else {
        type = atmid.substr(0,1);
      }
    } else if ( atmid[0] == ' ' ) {
      type = atmid.substr(1,1); // one char element
    } else {
      type = atmid.substr(1,2);
    }

    if ( atmid == resname ) {
      type = atmid;
      if ( type.size() == 2 ) type[1] = toupper(type[1]);
    } else if ( resname == "ADR" || resname == "COA" || resname == "FAD" ||
        resname == "GPG" || resname == "NAD" || resname == "NAL" ||
        resname == "NDP" || resname == "ABA" )  {
      if ( type.size() > 1 ) type = type.substr(0,1);
    } else if ( isdigit(type[0]) ) {
      type = type.substr(1,1);
    } else if ( type.size() > 1 && isdigit(type[1]) ) {
      type = type.substr(0,1);
    } else if ( type.size() > 1 && isalpha(type[1]) ) {
      if ( type[0] == 'O' && type[1] == 'H' ) {
        type = type.substr(0,1); // no "Oh" element (e.g. 1MBN)
      } else if ( islower(type[1]) ) {
        type[1] = toupper(type[1]);
      }
    }
  }
  return type;
}


// quick and dirty PDB read where we only care about heavyatom locations and atom ids
void
readPDBcoords(std::string filename, poseCoords &atmlist) {
  std::ifstream inpdb(filename.c_str());
  std::string buf;

  while ( std::getline(inpdb, buf ) ) {
    if ( buf.substr(0,4)!="ATOM" && buf.substr(0,6)!="HETATM" ) continue;
    poseCoord atom_i;

    atom_i.x_[0] = atof(buf.substr(30,8).c_str());
    atom_i.x_[1] = atof(buf.substr(38,8).c_str());
    atom_i.x_[2] = atof(buf.substr(46,8).c_str());
    atom_i.B_ = atof(buf.substr(60,6).c_str());

    atom_i.elt_ = name2elt( buf ); // horrible hacky logic mapping name->elt (could use PDB fields 76-77 if on by default)
    if ( atom_i.elt_ == "H" ) continue;

    atmlist.push_back( atom_i );
  }
}


///////////////////////////////////////////////////////////////////////////////
void
densityTools()
{
  using namespace protocols::moves;
  using namespace scoring;
  using namespace basic::options;
  using namespace basic::options::OptionKeys;

  // outputs
  Size nresobins = option[ denstools::nresbins ]();
  bool bin_squared =  option[ denstools::bin_squared ]();
  utility::vector1< core::Real > resobins, mapI, maskedmapI, modelI, maskI;
  utility::vector1< core::Size > resobin_counts;
  utility::vector1< core::Real > perResCC;

  utility::vector1< core::Real > mapmapFSC, maskedMapMapFSC;
  utility::vector1< core::Real > modelmapFSC, maskedModelMapFSC;

  // resolution limits for analysis
  core::Real hires = option[ denstools::hires ]();
  core::Real lowres = option[ denstools::lowres ]();
  core::Real truncate_hires = option[ denstools::truncate_hires ]();
  core::Real truncate_lowres = option[ denstools::truncate_lowres ]();
  core::Real mask_radius = option[ denstools::mask_radius ]();
  bool perres =option[ denstools::perres ]();

  ObjexxFCL::FArray3D< double > rhoC, rhoMask, rhoOmask, rhoO2mask;
  ObjexxFCL::FArray3D< std::complex<double> > FrhoC, FrhoMask, FrhoCmask, FrhoOmask, FrhoO2mask;
  ObjexxFCL::FArray3D< std::complex<double> > FrhoO, FrhoO2;

  if ( hires == 0.0 ) hires = core::scoring::electron_density::getDensityMap().maxNominalRes();
  if ( truncate_hires == 0.0 ) truncate_hires = core::scoring::electron_density::getDensityMap().maxNominalRes();

  runtime_assert( lowres > hires );
  runtime_assert( truncate_lowres > truncate_hires );

  hires = 1.0/hires;
  lowres = 1.0/lowres;
  truncate_lowres = 1.0/truncate_lowres;
  truncate_hires = 1.0/truncate_hires;

  //  fft
  std::cout << "Stage 1: FFT rho_obs" << std::endl;
  numeric::fourier::fft3(core::scoring::electron_density::getDensityMap().data(), FrhoO);
  core::scoring::electron_density::getDensityMap().getIntensities(FrhoO, nresobins, lowres, hires, mapI, bin_squared);

  // load model, mask density
  std::cout << "Stage 2: model processing" << std::endl;
  bool userpose = (option[ in::file::s ].user());

  poseCoords pose;
  core::pose::Pose fullpose;  // needed for per-residue stats (if requested)
  std::string pdbfile;
  if ( userpose && !option[ denstools::truncate_map ] ) {
    pdbfile = basic::options::start_file();
    readPDBcoords( pdbfile, pose );

    std::cout << "       : calc rho_c + FFT" << std::endl;
    core::scoring::electron_density::getDensityMap().calcRhoC( pose, mask_radius, rhoC, rhoMask );
    numeric::fourier::fft3(rhoC, FrhoC);
    numeric::fourier::fft3(rhoMask, FrhoMask);

    // apply mask
    std::cout << "       : mask map" << std::endl;
    rhoOmask = rhoMask;
    for ( int i=0; i<rhoC.u1()*rhoC.u2()*rhoC.u3(); ++i ) {
      rhoOmask[i] *= core::scoring::electron_density::getDensityMap().data()[i];
    }
    numeric::fourier::fft3(rhoOmask, FrhoOmask);

    // intensities
    std::cout << "       : get intensities" << std::endl;
    core::scoring::electron_density::getDensityMap().getIntensities(FrhoC, nresobins, lowres, hires, modelI, bin_squared);
    core::scoring::electron_density::getDensityMap().getIntensities(FrhoMask, nresobins, lowres, hires, maskI, bin_squared);
    core::scoring::electron_density::getDensityMap().getIntensities(FrhoOmask, nresobins, lowres, hires, maskedmapI, bin_squared);
  }

  core::scoring::electron_density::getDensityMap().getResolutionBins(nresobins, lowres, hires, resobins, resobin_counts, bin_squared);

  // rescale
  if ( option[ denstools::truncate_map ] ) {
    utility::vector1< core::Real > trunc = mapI;
    for ( int i=1; i<=(int)mapI.size(); ++i ) {
      if ( resobins[i] <= truncate_lowres || resobins[i] >= truncate_hires ) {
        trunc[i] = 0;
      } else {
        trunc[i] = 1;
      }
    }

    core::scoring::electron_density::getDensityMap().scaleIntensities( trunc, lowres, hires, bin_squared );
    core::scoring::electron_density::getDensityMap().writeMRC( "map_trunc.mrc" );
    std::cout << "Wrote truncated map" << std::endl;
    return;
  }

  // map vs map stats
  std::cout << "Stage 3: map vs map" << std::endl;
  core::scoring::electron_density::ElectronDensity mapAlt;
  bool usermap = option[ edensity::alt_mapfile ].user();
  if ( usermap ) {
    std::string mapfile = option[ edensity::alt_mapfile ];
    core::Real mapreso = option[ edensity::mapreso ]();
    core::Real mapsampling = option[ edensity::grid_spacing ]();
    std::cout << "       : load alt_map + FFT" << std::endl;
    mapAlt.readMRCandResize( mapfile , mapreso , mapsampling );
    numeric::fourier::fft3(mapAlt.data(), FrhoO2);

    if ( userpose ) {
      std::cout << "       : mask alt_map" << std::endl;
      rhoO2mask = rhoMask;
      for ( int i=0; i<rhoC.u1()*rhoC.u2()*rhoC.u3(); ++i ) {
        rhoO2mask[i] *= mapAlt.data()[i];
      }
      numeric::fourier::fft3(rhoO2mask, FrhoO2mask);
    }

    // unmasked
    std::cout << "       : get FSCs" << std::endl;
    core::scoring::electron_density::getDensityMap().getFSC(FrhoO, FrhoO2, nresobins, lowres, hires, mapmapFSC, bin_squared);

    // masked
    if ( userpose ) {
      core::scoring::electron_density::getDensityMap().getFSC(FrhoOmask, FrhoO2mask, nresobins, lowres, hires, mapmapFSC, bin_squared);
    }
  }

  // [3] model-map stats (intensity + model v map FSC + RSCC + per-res corrleations)
  Real modelMapFSCsum=0, maskedModelMapFSCsum=0, RSCC=0;

  std::cout << "Stage 4: model vs map" << std::endl;
  if ( userpose ) {
    if ( perres ) {
      std::cout << "       : per-res" << std::endl;
      core::chemical::ResidueTypeSetCAP rsd_set;
      rsd_set = core::chemical::ChemicalManager::get_instance()->residue_type_set( "fa_standard" );
      core::import_pose::pose_from_pdb( fullpose, pdbfile );

      core::Size nres = fullpose.total_residue();
      perResCC.resize( nres, 0.0 );

      protocols::electron_density::SetupForDensityScoringMoverOP dockindens( new protocols::electron_density::SetupForDensityScoringMover );
      dockindens->apply( fullpose );

      core::scoring::electron_density::getDensityMap().set_nres( nres );
      core::scoring::electron_density::getDensityMap().setScoreWindowContext( true );
      core::scoring::electron_density::getDensityMap().setWindow( option[ denstools::rscc_wdw ]() );

      core::scoring::ScoreFunctionOP scorefxn = core::scoring::get_score_function();
      scorefxn->set_weight( core::scoring::elec_dens_window, 1.0 );
      (*scorefxn)(fullpose);

      for ( core::uint r = 1; r <= nres; ++r ) {
        perResCC[r] = core::scoring::electron_density::getDensityMap().matchRes( r , fullpose.residue(r), fullpose, NULL , false);
      }
    }

    std::cout << "       : FSCs" << std::endl;
    core::scoring::electron_density::getDensityMap().getFSC(FrhoO, FrhoC, nresobins, lowres, hires, modelmapFSC, bin_squared);
    core::scoring::electron_density::getDensityMap().getFSC(FrhoOmask, FrhoC, nresobins, lowres, hires, maskedModelMapFSC, bin_squared);

    std::cout << "       : RSCC" << std::endl;
    RSCC = core::scoring::electron_density::getDensityMap().getRSCC( rhoC, rhoMask );

    // sum FSC over reso bins
    for ( Size i=1; i<=resobins.size(); ++i ) {
      modelMapFSCsum += modelmapFSC[i];
      maskedModelMapFSCsum += maskedModelMapFSC[i];
    }
    modelMapFSCsum /= resobins.size();
    maskedModelMapFSCsum /= resobins.size();
  }

  // verbose
  if ( option[ denstools::verbose ]() ) {
    std::cerr << "res count";
    if ( mapI.size() ) std::cerr << " I_map";
    if ( maskedmapI.size() ) std::cerr << " I_map_masked";
    if ( modelI.size() ) std::cerr << " I_model" ;
    if ( maskI.size() ) std::cerr << " I_mask" ;
    if ( mapmapFSC.size() )  std::cerr << " FSC_mapmap";
    if ( modelmapFSC.size() )  std::cerr << " FSC_modelmap";
    if ( maskedModelMapFSC.size() )  std::cerr << " FSC_maskedmodelmap";
    std::cerr << std::endl;
    for ( Size i=1; i<=resobins.size(); ++i ) {
      std::cerr << resobins[i] << " " << resobin_counts[i];
      if ( mapI.size() ) std::cerr << " " << mapI[i];
      if ( maskedmapI.size() ) std::cerr << " " << maskedmapI[i];
      if ( modelI.size() ) std::cerr << " " << modelI[i];
      if ( maskI.size() ) std::cerr << " " << maskI[i];
      if ( mapmapFSC.size() )  std::cerr << " " << mapmapFSC[i];
      if ( modelmapFSC.size() )  std::cerr << " " << modelmapFSC[i];
      if ( maskedModelMapFSC.size() )  std::cerr << " " << maskedModelMapFSC[i];
      std::cerr << std::endl;
    }
  }

  // dump
  if ( userpose && option[ denstools::dump_map_and_mask ]() ) {
    core::scoring::electron_density::getDensityMap().set_data(rhoOmask);
    core::scoring::electron_density::getDensityMap().writeMRC( "map_masked.mrc" );

    core::scoring::electron_density::getDensityMap().set_data(rhoC);
    core::scoring::electron_density::getDensityMap().writeMRC( "model_dens.mrc" );

    core::scoring::electron_density::getDensityMap().set_data(rhoMask);
    core::scoring::electron_density::getDensityMap().writeMRC( "model_mask.mrc" );
  }

  if ( userpose && option[ denstools::perres ]() ) {
    for ( core::uint r = 1; r <= perResCC.size(); ++r ) {
      std::cerr << "PERRESCC residue " << r << " " << perResCC[r] << std::endl;
    }
  }

  // compact
  if ( userpose ) {
    std::cerr << pdbfile << "RSCC/FSC/FSCmask: " << RSCC << " " << modelMapFSCsum << " " << maskedModelMapFSCsum;
    std::cerr << std::endl;
  }
}


///////////////////////////////////////////////////////////////////////////////

int
main( int argc, char * argv [] )
{
  try {
    // options, random initialization
    NEW_OPT(denstools::lowres, "low res limit", 1000.0);
    NEW_OPT(denstools::hires, "high res limit", 0.0);
    NEW_OPT(edensity::alt_mapfile, "alt mapfile", "");
    NEW_OPT(denstools::nresbins, "# resolution bins for statistics", 200);
    NEW_OPT(denstools::truncate_map, "dump map with map I scaled to to model I?", false);
    NEW_OPT(denstools::truncate_lowres, "low res truncation", 1000.0);
    NEW_OPT(denstools::truncate_hires, "high res truncation", 0.0);
    NEW_OPT(denstools::dump_map_and_mask, "dump mrc of rho_calc and eps_calc", false);
    NEW_OPT(denstools::mask, "mask all calcs", false);
    NEW_OPT(denstools::mask_modelmap, "mask model-map calcs only", false);
    NEW_OPT(denstools::mask_radius, "radius for masking", 0.0);
    NEW_OPT(denstools::perres, "output per-residue stats", false);
    NEW_OPT(denstools::verbose, "dump extra output", false);
    NEW_OPT(denstools::super_verbose, "dump a lot of extra output", false);
    NEW_OPT(denstools::bin_squared, "bin uniformly in 1/res^2 (default bins 1/res)", false);
    NEW_OPT(denstools::rscc_wdw, "sliding window to calculate rscc", 3);

    devel::init( argc, argv );
    densityTools();
  } catch ( utility::excn::EXCN_Base const & e ) {
    std::cout << "caught exception " << e.msg() << std::endl;
    return -1;
  }
  return 0;
}