2 * This file is part of the GROMACS molecular simulation package.
4 * Copyright (c) 1991-2000, University of Groningen, The Netherlands.
5 * Copyright (c) 2001-2006, The GROMACS development team.
6 * Copyright (c) 2008,2009,2010,2011,2012,2013,2014,2015,2016,2017,2018, by the GROMACS development team, led by
7 * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl,
8 * and including many others, as listed in the AUTHORS file in the
9 * top-level source directory and at http://www.gromacs.org.
11 * GROMACS is free software; you can redistribute it and/or
12 * modify it under the terms of the GNU Lesser General Public License
13 * as published by the Free Software Foundation; either version 2.1
14 * of the License, or (at your option) any later version.
16 * GROMACS is distributed in the hope that it will be useful,
17 * but WITHOUT ANY WARRANTY; without even the implied warranty of
18 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
19 * Lesser General Public License for more details.
21 * You should have received a copy of the GNU Lesser General Public
22 * License along with GROMACS; if not, see
23 * http://www.gnu.org/licenses, or write to the Free Software Foundation,
24 * Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
26 * If you want to redistribute modifications to GROMACS, please
27 * consider that scientific software is very special. Version
28 * control is crucial - bugs must be traceable. We will be happy to
29 * consider code for inclusion in the official distribution, but
30 * derived work must not be called official GROMACS. Details are found
31 * in the README & COPYING files - if they are missing, get the
32 * official version at http://www.gromacs.org.
34 * To help us fund GROMACS development, we humbly ask that you cite
35 * the research papers on the package. Check out http://www.gromacs.org.
39 * Implements gmx::analysismodules::Sasa.
41 * \author Teemu Murtola <teemu.murtola@gmail.com> (C++ conversion)
42 * \ingroup module_trajectoryanalysis
52 #include "gromacs/analysisdata/analysisdata.h"
53 #include "gromacs/analysisdata/modules/average.h"
54 #include "gromacs/analysisdata/modules/plot.h"
55 #include "gromacs/fileio/confio.h"
56 #include "gromacs/fileio/pdbio.h"
57 #include "gromacs/math/units.h"
58 #include "gromacs/math/vec.h"
59 #include "gromacs/options/basicoptions.h"
60 #include "gromacs/options/filenameoption.h"
61 #include "gromacs/options/ioptionscontainer.h"
62 #include "gromacs/pbcutil/pbc.h"
63 #include "gromacs/selection/selection.h"
64 #include "gromacs/selection/selectionoption.h"
65 #include "gromacs/topology/atomprop.h"
66 #include "gromacs/topology/symtab.h"
67 #include "gromacs/topology/topology.h"
68 #include "gromacs/trajectory/trajectoryframe.h"
69 #include "gromacs/trajectoryanalysis/analysismodule.h"
70 #include "gromacs/trajectoryanalysis/analysissettings.h"
71 #include "gromacs/trajectoryanalysis/topologyinformation.h"
72 #include "gromacs/utility/exceptions.h"
73 #include "gromacs/utility/futil.h"
74 #include "gromacs/utility/pleasecite.h"
75 #include "gromacs/utility/smalloc.h"
76 #include "gromacs/utility/stringutil.h"
77 #include "gromacs/utility/unique_cptr.h"
79 #include "surfacearea.h"
84 namespace analysismodules
90 //! \addtogroup module_trajectoryanalysis
93 //! Tracks information on two nearest neighbors of a single surface dot.
96 //! Index of the second nearest neighbor dot.
98 //! Index of the nearest neighbor dot.
100 //! Squared distance to `aa`.
102 //! Squared distance to `ab`.
107 * Updates nearest neighbor information for a surface dot.
109 * \param[in,out] c Nearest neighbor information array to update.
110 * \param[in] i Index in `c` to update.
111 * \param[in] j Index of the other surface dot to add to the array.
112 * \param[in] d2 Squared distance between `i` and `j`.
114 void add_rec(t_conect c[], int i, int j, real d2)
121 else if (c[i].ab == -1)
126 else if (d2 < c[i].d2a)
131 else if (d2 < c[i].d2b)
136 /* Swap them if necessary: a must be larger than b */
137 if (c[i].d2a < c[i].d2b)
149 * Adds CONECT records for surface dots.
151 * \param[in] fn PDB file to append the CONECT records to.
152 * \param[in] n Number of dots in `x`.
153 * \param[in] x Array of surface dot positions.
155 * Adds a CONECT record that connects each surface dot to its two nearest
156 * neighbors. The function is copied verbatim from the old gmx_sas.c
159 void do_conect(const char *fn, int n, rvec x[])
167 fprintf(stderr, "Building CONECT records\n");
169 for (i = 0; (i < n); i++)
171 c[i].aa = c[i].ab = -1;
174 for (i = 0; (i < n); i++)
176 for (j = i+1; (j < n); j++)
178 rvec_sub(x[i], x[j], dx);
180 add_rec(c, i, j, d2);
181 add_rec(c, j, i, d2);
184 fp = gmx_ffopen(fn, "a");
185 for (i = 0; (i < n); i++)
187 if ((c[i].aa == -1) || (c[i].ab == -1))
189 fprintf(stderr, "Warning dot %d has no connections\n", i+1);
191 fprintf(fp, "CONECT%5d%5d%5d\n", i+1, c[i].aa+1, c[i].ab+1);
198 * Plots the surface into a PDB file, optionally including the original atoms.
200 void connolly_plot(const char *fn, int ndots, const real dots[], rvec x[], t_atoms *atoms,
201 t_symtab *symtab, int ePBC, const matrix box, gmx_bool bIncludeSolute)
203 const char *const atomnm = "DOT";
204 const char *const resnm = "DOT";
205 const char *const title = "Connolly Dot Surface Generated by gmx sasa";
207 int i, i0, r0, ii0, k;
215 srenew(atoms->atom, atoms->nr+ndots);
216 memset(&atoms->atom[i0], 0, sizeof(*atoms->atom)*ndots);
217 srenew(atoms->atomname, atoms->nr+ndots);
218 srenew(atoms->resinfo, r0+1);
219 atoms->atom[i0].resind = r0;
220 t_atoms_set_resinfo(atoms, i0, symtab, resnm, r0+1, ' ', 0, ' ');
221 if (atoms->pdbinfo != nullptr)
223 srenew(atoms->pdbinfo, atoms->nr+ndots);
225 snew(xnew, atoms->nr+ndots);
226 for (i = 0; (i < atoms->nr); i++)
228 copy_rvec(x[i], xnew[i]);
230 for (i = k = 0; (i < ndots); i++)
233 atoms->atomname[ii0] = put_symtab(symtab, atomnm);
234 atoms->atom[ii0].resind = r0;
235 xnew[ii0][XX] = dots[k++];
236 xnew[ii0][YY] = dots[k++];
237 xnew[ii0][ZZ] = dots[k++];
238 if (atoms->pdbinfo != nullptr)
240 atoms->pdbinfo[ii0].type = epdbATOM;
241 atoms->pdbinfo[ii0].atomnr = ii0;
242 atoms->pdbinfo[ii0].bfac = 0.0;
243 atoms->pdbinfo[ii0].occup = 0.0;
246 atoms->nr = i0+ndots;
248 write_sto_conf(fn, title, atoms, xnew, nullptr, ePBC, const_cast<rvec *>(box));
254 init_t_atoms(&aaa, ndots, TRUE);
255 aaa.atom[0].resind = 0;
256 t_atoms_set_resinfo(&aaa, 0, symtab, resnm, 1, ' ', 0, ' ');
258 for (i = k = 0; (i < ndots); i++)
261 aaa.atomname[ii0] = put_symtab(symtab, atomnm);
262 aaa.pdbinfo[ii0].type = epdbATOM;
263 aaa.pdbinfo[ii0].atomnr = ii0;
264 aaa.atom[ii0].resind = 0;
265 xnew[ii0][XX] = dots[k++];
266 xnew[ii0][YY] = dots[k++];
267 xnew[ii0][ZZ] = dots[k++];
268 aaa.pdbinfo[ii0].bfac = 0.0;
269 aaa.pdbinfo[ii0].occup = 0.0;
272 write_sto_conf(fn, title, &aaa, xnew, nullptr, ePBC, const_cast<rvec *>(box));
273 do_conect(fn, ndots, xnew);
279 /********************************************************************
280 * Actual analysis module
284 * Implements `gmx sas` trajectory analysis module.
286 class Sasa : public TrajectoryAnalysisModule
291 void initOptions(IOptionsContainer *options,
292 TrajectoryAnalysisSettings *settings) override;
293 void initAnalysis(const TrajectoryAnalysisSettings &settings,
294 const TopologyInformation &top) override;
296 TrajectoryAnalysisModuleDataPointer startFrames(
297 const AnalysisDataParallelOptions &opt,
298 const SelectionCollection &selections) override;
299 void analyzeFrame(int frnr, const t_trxframe &fr, t_pbc *pbc,
300 TrajectoryAnalysisModuleData *pdata) override;
302 void finishAnalysis(int nframes) override;
303 void writeOutput() override;
307 * Surface areas as a function of time.
309 * First column is for the calculation group, and the rest for the
310 * output groups. This data is always produced.
314 * Per-atom surface areas as a function of time.
316 * Contains one data set for each column in `area_`.
317 * Each column corresponds to a selection position in `surfaceSel_`.
318 * This data is only produced if atom or residue areas have been
321 AnalysisData atomArea_;
323 * Per-residue surface areas as a function of time.
325 * Contains one data set for each column in `area_`.
326 * Each column corresponds to a distinct residue `surfaceSel_`.
327 * For example, if `surfaceSel_` selects residues 2, 5, and 7, there
328 * will be three columns here.
329 * This data is only produced if atom or residue areas have been
332 AnalysisData residueArea_;
334 * Free energy estimates as a function of time.
336 * Column layout is the same as for `area_`.
337 * This data is only produced if the output is requested.
339 AnalysisData dgSolv_;
341 * Total volume and density of the calculation group as a function of
344 * The first column is the volume and the second column is the density.
345 * This data is only produced if the output is requested.
347 AnalysisData volume_;
350 * The selection to calculate the surface for.
352 * Selection::originalId() and Selection::mappedId() store the mapping
353 * from the positions to the columns of `residueArea_`.
354 * The selection is computed with SelectionOption::dynamicMask(), i.e.,
355 * even in the presence of a dynamic selection, the number of returned
356 * positions is fixed, and SelectionPosition::selected() is used.
358 Selection surfaceSel_;
360 * List of optional additional output groups.
362 * Each of these must be a subset of the `surfaceSel_`.
363 * Selection::originalId() and Selection::mappedId() store the mapping
364 * from the positions to the corresponsing positions in `surfaceSel_`.
366 SelectionList outputSel_;
369 std::string fnAtomArea_;
370 std::string fnResidueArea_;
371 std::string fnDGSolv_;
372 std::string fnVolume_;
373 std::string fnConnolly_;
379 bool bIncludeSolute_;
382 //! Combined VdW and probe radii for each atom in the calculation group.
383 std::vector<real> radii_;
385 * Solvation free energy coefficients for each atom in the calculation
388 * Empty if the free energy output has not been requested.
390 std::vector<real> dgsFactor_;
391 //! Calculation algorithm.
392 SurfaceAreaCalculator calculator_;
394 // Copy and assign disallowed by base.
398 : solsize_(0.14), ndots_(24), dgsDefault_(0), bIncludeSolute_(true), top_(nullptr)
401 registerAnalysisDataset(&area_, "area");
402 registerAnalysisDataset(&atomArea_, "atomarea");
403 registerAnalysisDataset(&residueArea_, "resarea");
404 registerAnalysisDataset(&dgSolv_, "dgsolv");
405 registerAnalysisDataset(&volume_, "volume");
409 Sasa::initOptions(IOptionsContainer *options, TrajectoryAnalysisSettings *settings)
411 static const char *const desc[] = {
412 "[THISMODULE] computes solvent accessible surface areas.",
413 "See Eisenhaber F, Lijnzaad P, Argos P, Sander C, & Scharf M",
414 "(1995) J. Comput. Chem. 16, 273-284 for the algorithm used.",
415 "With [TT]-q[tt], the Connolly surface can be generated as well",
416 "in a [REF].pdb[ref] file where the nodes are represented as atoms",
417 "and the edges connecting the nearest nodes as CONECT records.",
418 "[TT]-odg[tt] allows for estimation of solvation free energies",
419 "from per-atom solvation energies per exposed surface area.[PAR]",
421 "The program requires a selection for the surface calculation to be",
422 "specified with [TT]-surface[tt]. This should always consist of all",
423 "non-solvent atoms in the system. The area of this group is always",
424 "calculated. Optionally, [TT]-output[tt] can specify additional",
425 "selections, which should be subsets of the calculation group.",
426 "The solvent-accessible areas for these groups are also extracted",
427 "from the full surface.[PAR]",
429 "The average and standard deviation of the area over the trajectory",
430 "can be calculated per residue and atom (options [TT]-or[tt] and",
431 "[TT]-oa[tt]).[PAR]",
432 //"In combination with the latter option an [REF].itp[ref] file can be",
433 //"generated (option [TT]-i[tt])",
434 //"which can be used to restrain surface atoms.[PAR]",
436 "With the [TT]-tv[tt] option the total volume and density of the",
437 "molecule can be computed. With [TT]-pbc[tt] (the default), you",
438 "must ensure that your molecule/surface group is not split across PBC.",
439 "Otherwise, you will get non-sensical results.",
440 "Please also consider whether the normal probe radius is appropriate",
441 "in this case or whether you would rather use, e.g., 0. It is good",
442 "to keep in mind that the results for volume and density are very",
443 "approximate. For example, in ice Ih, one can easily fit water molecules in the",
444 "pores which would yield a volume that is too low, and surface area and density",
445 "that are both too high."
448 settings->setHelpText(desc);
450 options->addOption(FileNameOption("o").filetype(eftPlot).outputFile().required()
451 .store(&fnArea_).defaultBasename("area")
452 .description("Total area as a function of time"));
453 options->addOption(FileNameOption("odg").filetype(eftPlot).outputFile()
454 .store(&fnDGSolv_).defaultBasename("dgsolv")
455 .description("Estimated solvation free energy as a function of time"));
456 options->addOption(FileNameOption("or").filetype(eftPlot).outputFile()
457 .store(&fnResidueArea_).defaultBasename("resarea")
458 .description("Average area per residue"));
459 options->addOption(FileNameOption("oa").filetype(eftPlot).outputFile()
460 .store(&fnAtomArea_).defaultBasename("atomarea")
461 .description("Average area per atom"));
462 options->addOption(FileNameOption("tv").filetype(eftPlot).outputFile()
463 .store(&fnVolume_).defaultBasename("volume")
464 .description("Total volume and density as a function of time"));
465 options->addOption(FileNameOption("q").filetype(eftPDB).outputFile()
466 .store(&fnConnolly_).defaultBasename("connolly")
467 .description("PDB file for Connolly surface"));
468 //options->addOption(FileNameOption("i").filetype(eftITP).outputFile()
469 // .store(&fnRestraints_).defaultBasename("surfat")
470 // .description("Topology file for position restraings on surface atoms"));
473 options->addOption(DoubleOption("probe").store(&solsize_)
474 .description("Radius of the solvent probe (nm)"));
475 options->addOption(IntegerOption("ndots").store(&ndots_)
476 .description("Number of dots per sphere, more dots means more accuracy"));
477 //options->addOption(DoubleOption("minarea").store(&minarea_)
478 // .description("The minimum area (nm^2) to count an atom as a surface atom when writing a position restraint file (see help)"));
479 options->addOption(BooleanOption("prot").store(&bIncludeSolute_)
480 .description("Output the protein to the Connolly [REF].pdb[ref] file too"));
481 options->addOption(DoubleOption("dgs").store(&dgsDefault_)
482 .description("Default value for solvation free energy per area (kJ/mol/nm^2)"));
484 // Selections must select atoms for the VdW radii lookup to work.
485 // The calculation group uses dynamicMask() so that the coordinates
486 // match a static array of VdW radii.
487 options->addOption(SelectionOption("surface").store(&surfaceSel_)
488 .required().onlySortedAtoms().dynamicMask()
489 .description("Surface calculation selection"));
490 options->addOption(SelectionOption("output").storeVector(&outputSel_)
491 .onlySortedAtoms().multiValue()
492 .description("Output selection(s)"));
494 // Atom names etc. are required for the VdW radii lookup.
495 settings->setFlag(TrajectoryAnalysisSettings::efRequireTop);
499 Sasa::initAnalysis(const TrajectoryAnalysisSettings &settings,
500 const TopologyInformation &top)
502 const t_atoms &atoms = top.topology()->atoms;
503 top_ = top.topology();
505 //bITP = opt2bSet("-i", nfile, fnm);
507 !fnResidueArea_.empty() || !fnAtomArea_.empty(); // || bITP;
508 const bool bDGsol = !fnDGSolv_.empty();
513 fprintf(stderr, "Probe size too small, setting it to %g\n", solsize_);
518 fprintf(stderr, "Ndots too small, setting it to %d\n", ndots_);
521 please_cite(stderr, "Eisenhaber95");
522 //if ((top.ePBC() != epbcXYZ) || (TRICLINIC(fr.box)))
524 // fprintf(stderr, "\n\nWARNING: non-rectangular boxes may give erroneous results or crashes.\n"
525 // "Analysis based on vacuum simulations (with the possibility of evaporation)\n"
526 // "will certainly crash the analysis.\n\n");
531 if (!top.hasFullTopology())
533 GMX_THROW(InconsistentInputError("Cannot compute Delta G of solvation without a tpr file"));
537 if (strcmp(*(atoms.atomtype[0]), "?") == 0)
539 GMX_THROW(InconsistentInputError("Your input tpr file is too old (does not contain atom types). Cannot not compute Delta G of solvation"));
543 printf("Free energy of solvation predictions:\n");
544 please_cite(stdout, "Eisenberg86a");
549 // Now compute atomic radii including solvent probe size.
550 // Also, fetch solvation free energy coefficients and
551 // compute the residue indices that map the calculation atoms
552 // to the columns of residueArea_.
553 radii_.reserve(surfaceSel_.posCount());
556 dgsFactor_.reserve(surfaceSel_.posCount());
559 const int resCount = surfaceSel_.initOriginalIdsToGroup(top.mtop(), INDEX_RES);
561 // TODO: Not exception-safe, but nice solution would be to have a C++
562 // atom properties class...
563 gmx_atomprop_t aps = gmx_atomprop_init();
565 ArrayRef<const int> atomIndices = surfaceSel_.atomIndices();
567 for (int i = 0; i < surfaceSel_.posCount(); i++)
569 const int ii = atomIndices[i];
570 const int resind = atoms.atom[ii].resind;
572 if (!gmx_atomprop_query(aps, epropVDW,
573 *(atoms.resinfo[resind].name),
574 *(atoms.atomname[ii]), &radius))
578 radii_.push_back(radius + solsize_);
582 if (!gmx_atomprop_query(aps, epropDGsol,
583 *(atoms.resinfo[resind].name),
584 *(atoms.atomtype[ii]), &dgsFactor))
586 dgsFactor = dgsDefault_;
588 dgsFactor_.push_back(dgsFactor);
593 fprintf(stderr, "WARNING: could not find a Van der Waals radius for %d atoms\n", ndefault);
595 gmx_atomprop_destroy(aps);
597 // Pre-compute mapping from the output groups to the calculation group,
598 // and store it in the selection ID map for easy lookup.
599 for (size_t g = 0; g < outputSel_.size(); ++g)
601 ArrayRef<const int> outputIndices = outputSel_[g].atomIndices();
602 for (int i = 0, j = 0; i < outputSel_[g].posCount(); ++i)
604 while (j < surfaceSel_.posCount() && outputIndices[i] > atomIndices[j])
608 if (j == surfaceSel_.posCount() || outputIndices[i] != atomIndices[j])
610 const std::string message
611 = formatString("Output selection '%s' is not a subset of "
612 "the surface selection (atom %d is the first "
613 "atom not in the surface selection)",
614 outputSel_[g].name(), outputIndices[i] + 1);
615 GMX_THROW(InconsistentInputError(message));
617 outputSel_[g].setOriginalId(i, j);
621 calculator_.setDotCount(ndots_);
622 calculator_.setRadii(radii_);
624 // Initialize all the output data objects and initialize the output plotters.
626 area_.setColumnCount(0, 1 + outputSel_.size());
628 AnalysisDataPlotModulePointer plotm(
629 new AnalysisDataPlotModule(settings.plotSettings()));
630 plotm->setFileName(fnArea_);
631 plotm->setTitle("Solvent Accessible Surface");
632 plotm->setXAxisIsTime();
633 plotm->setYLabel("Area (nm\\S2\\N)");
634 plotm->appendLegend("Total");
635 for (size_t i = 0; i < outputSel_.size(); ++i)
637 plotm->appendLegend(outputSel_[i].name());
639 area_.addModule(plotm);
644 atomArea_.setDataSetCount(1 + outputSel_.size());
645 residueArea_.setDataSetCount(1 + outputSel_.size());
646 for (size_t i = 0; i <= outputSel_.size(); ++i)
648 atomArea_.setColumnCount(i, surfaceSel_.posCount());
649 residueArea_.setColumnCount(i, resCount);
652 AnalysisDataAverageModulePointer avem(new AnalysisDataAverageModule);
653 for (int i = 0; i < surfaceSel_.posCount(); ++i)
655 avem->setXAxisValue(i, surfaceSel_.position(i).atomIndices()[0] + 1);
657 atomArea_.addModule(avem);
658 if (!fnAtomArea_.empty())
660 AnalysisDataPlotModulePointer plotm(
661 new AnalysisDataPlotModule(settings.plotSettings()));
662 plotm->setFileName(fnAtomArea_);
663 plotm->setTitle("Area per atom over the trajectory");
664 plotm->setXLabel("Atom");
665 plotm->setXFormat(8, 0);
666 plotm->setYLabel("Area (nm\\S2\\N)");
667 plotm->setErrorsAsSeparateColumn(true);
668 plotm->appendLegend("Average (nm\\S2\\N)");
669 plotm->appendLegend("Standard deviation (nm\\S2\\N)");
670 avem->addModule(plotm);
674 AnalysisDataAverageModulePointer avem(new AnalysisDataAverageModule);
676 for (int i = 0; i < surfaceSel_.posCount(); ++i)
678 const int residueGroup = surfaceSel_.position(i).mappedId();
679 if (residueGroup >= nextRow)
681 GMX_ASSERT(residueGroup == nextRow,
682 "Inconsistent (non-uniformly increasing) residue grouping");
683 const int atomIndex = surfaceSel_.position(i).atomIndices()[0];
684 const int residueIndex = atoms.atom[atomIndex].resind;
685 avem->setXAxisValue(nextRow, atoms.resinfo[residueIndex].nr);
689 residueArea_.addModule(avem);
690 if (!fnResidueArea_.empty())
692 AnalysisDataPlotModulePointer plotm(
693 new AnalysisDataPlotModule(settings.plotSettings()));
694 plotm->setFileName(fnResidueArea_);
695 plotm->setTitle("Area per residue over the trajectory");
696 plotm->setXLabel("Residue");
697 plotm->setXFormat(8, 0);
698 plotm->setYLabel("Area (nm\\S2\\N)");
699 plotm->setErrorsAsSeparateColumn(true);
700 plotm->appendLegend("Average (nm\\S2\\N)");
701 plotm->appendLegend("Standard deviation (nm\\S2\\N)");
702 avem->addModule(plotm);
707 if (!fnDGSolv_.empty())
709 dgSolv_.setColumnCount(0, 1 + outputSel_.size());
710 AnalysisDataPlotModulePointer plotm(
711 new AnalysisDataPlotModule(settings.plotSettings()));
712 plotm->setFileName(fnDGSolv_);
713 plotm->setTitle("Free Energy of Solvation");
714 plotm->setXAxisIsTime();
715 plotm->setYLabel("D Gsolv");
716 plotm->appendLegend("Total");
717 for (size_t i = 0; i < outputSel_.size(); ++i)
719 plotm->appendLegend(outputSel_[i].name());
721 dgSolv_.addModule(plotm);
724 if (!fnVolume_.empty())
726 volume_.setColumnCount(0, 2);
727 AnalysisDataPlotModulePointer plotm(
728 new AnalysisDataPlotModule(settings.plotSettings()));
729 plotm->setFileName(fnVolume_);
730 plotm->setTitle("Volume and Density");
731 plotm->setXAxisIsTime();
732 plotm->appendLegend("Volume (nm\\S3\\N)");
733 plotm->appendLegend("Density (g/l)");
734 volume_.addModule(plotm);
739 * Temporary memory for use within a single-frame calculation.
741 class SasaModuleData : public TrajectoryAnalysisModuleData
745 * Reserves memory for the frame-local data.
747 * `residueCount` will be zero if per-residue data is not being
750 SasaModuleData(TrajectoryAnalysisModule *module,
751 const AnalysisDataParallelOptions &opt,
752 const SelectionCollection &selections,
753 int atomCount, int residueCount)
754 : TrajectoryAnalysisModuleData(module, opt, selections)
756 index_.reserve(atomCount);
757 // If the calculation group is not dynamic, pre-calculate
758 // the index, since it is not going to change.
759 for (int i = 0; i < atomCount; ++i)
763 atomAreas_.resize(atomCount);
764 res_a_.resize(residueCount);
767 void finish() override { finishDataHandles(); }
769 //! Indices of the calculation selection positions selected for the frame.
770 std::vector<int> index_;
772 * Atom areas for each calculation selection position for the frame.
774 * One entry for each position in the calculation group.
775 * Values for atoms not selected are set to zero.
777 std::vector<real> atomAreas_;
779 * Working array to accumulate areas for each residue.
781 * One entry for each distinct residue in the calculation group;
782 * indices are not directly residue numbers or residue indices.
784 * This vector is empty if residue area calculations are not being
787 std::vector<real> res_a_;
790 TrajectoryAnalysisModuleDataPointer Sasa::startFrames(
791 const AnalysisDataParallelOptions &opt,
792 const SelectionCollection &selections)
794 return TrajectoryAnalysisModuleDataPointer(
795 new SasaModuleData(this, opt, selections, surfaceSel_.posCount(),
796 residueArea_.columnCount(0)));
800 * Helper method to compute the areas for a single selection.
802 * \param[in] surfaceSel The calculation selection.
803 * \param[in] sel The selection to compute the areas for (can be
804 * `surfaceSel` or one of the output selections).
805 * \param[in] atomAreas Atom areas for each position in `surfaceSel`.
806 * \param[in] dgsFactor Free energy coefficients for each position in
807 * `surfaceSel`. If empty, free energies are not calculated.
808 * \param[out] totalAreaOut Total area of `sel` (sum of atom areas it selects).
809 * \param[out] dgsolvOut Solvation free energy.
810 * Will be zero of `dgsFactor` is empty.
811 * \param atomAreaHandle Data handle to use for storing atom areas for `sel`.
812 * \param resAreaHandle Data handle to use for storing residue areas for `sel`.
813 * \param resAreaWork Work array for accumulating the residue areas.
814 * If empty, atom and residue areas are not calculated.
816 * `atomAreaHandle` and `resAreaHandle` are not used if `resAreaWork` is empty.
818 void computeAreas(const Selection &surfaceSel, const Selection &sel,
819 const std::vector<real> &atomAreas,
820 const std::vector<real> &dgsFactor,
821 real *totalAreaOut, real *dgsolvOut,
822 AnalysisDataHandle atomAreaHandle,
823 AnalysisDataHandle resAreaHandle,
824 std::vector<real> *resAreaWork)
826 const bool bResAt = !resAreaWork->empty();
827 const bool bDGsolv = !dgsFactor.empty();
833 std::fill(resAreaWork->begin(), resAreaWork->end(), 0.0_real);
835 for (int i = 0; i < sel.posCount(); ++i)
837 // Get the index of the atom in the calculation group.
838 // For the output groups, the mapping has been precalculated in
840 const int ii = (sel != surfaceSel ? sel.position(i).mappedId() : i);
841 if (!surfaceSel.position(ii).selected())
843 // For the calculation group, skip unselected atoms.
844 if (sel == surfaceSel)
848 GMX_THROW(InconsistentInputError("Output selection is not a subset of the surface selection"));
850 // Get the internal index of the matching residue.
851 // These have been precalculated in initAnalysis().
852 const int ri = surfaceSel.position(ii).mappedId();
853 const real atomArea = atomAreas[ii];
854 totalArea += atomArea;
857 atomAreaHandle.setPoint(ii, atomArea);
858 (*resAreaWork)[ri] += atomArea;
862 dgsolv += atomArea * dgsFactor[ii];
867 for (size_t i = 0; i < (*resAreaWork).size(); ++i)
869 resAreaHandle.setPoint(i, (*resAreaWork)[i]);
872 *totalAreaOut = totalArea;
877 Sasa::analyzeFrame(int frnr, const t_trxframe &fr, t_pbc *pbc,
878 TrajectoryAnalysisModuleData *pdata)
880 AnalysisDataHandle ah = pdata->dataHandle(area_);
881 AnalysisDataHandle dgh = pdata->dataHandle(dgSolv_);
882 AnalysisDataHandle aah = pdata->dataHandle(atomArea_);
883 AnalysisDataHandle rah = pdata->dataHandle(residueArea_);
884 AnalysisDataHandle vh = pdata->dataHandle(volume_);
885 const Selection &surfaceSel = pdata->parallelSelection(surfaceSel_);
886 const SelectionList &outputSel = pdata->parallelSelections(outputSel_);
887 SasaModuleData &frameData = *static_cast<SasaModuleData *>(pdata);
889 const bool bResAt = !frameData.res_a_.empty();
890 const bool bDGsol = !dgsFactor_.empty();
891 const bool bConnolly = (frnr == 0 && !fnConnolly_.empty());
893 // Update indices of selected atoms in the work array.
894 if (surfaceSel.isDynamic())
896 frameData.index_.clear();
897 for (int i = 0; i < surfaceSel.posCount(); ++i)
899 if (surfaceSel.position(i).selected())
901 frameData.index_.push_back(i);
906 // Determine what needs to be calculated.
908 if (bResAt || bDGsol || !outputSel.empty())
910 flag |= FLAG_ATOM_AREA;
916 if (volume_.columnCount() > 0)
921 // Do the low-level calculation.
922 // totarea and totvolume receive the values for the calculation group.
923 // area array contains the per-atom areas for the selected positions.
924 // surfacedots contains nsurfacedots entries, and contains the actual
926 real totarea, totvolume;
927 real *area = nullptr, *surfacedots = nullptr;
929 calculator_.calculate(surfaceSel.coordinates().data(), pbc,
930 frameData.index_.size(), frameData.index_.data(), flag,
931 &totarea, &totvolume, &area,
932 &surfacedots, &nsurfacedots);
933 // Unpack the atomwise areas into the frameData.atomAreas_ array for easier
934 // indexing in the case of dynamic surfaceSel.
937 if (surfaceSel.isDynamic())
939 std::fill(frameData.atomAreas_.begin(), frameData.atomAreas_.end(),
941 for (size_t i = 0; i < frameData.index_.size(); ++i)
943 frameData.atomAreas_[frameData.index_[i]] = area[i];
948 std::copy(area, area + surfaceSel.posCount(),
949 frameData.atomAreas_.begin());
953 const sfree_guard dotsGuard(surfacedots);
957 if (fr.natoms != top_->atoms.nr)
959 GMX_THROW(InconsistentInputError("Connolly plot (-q) is only supported for trajectories that contain all the atoms"));
961 // This is somewhat nasty, as it modifies the atoms and symtab
962 // structures. But since it is only used in the first frame, and no
963 // one else uses the topology after initialization, it may just work
964 // even with future parallelization.
965 connolly_plot(fnConnolly_.c_str(),
966 nsurfacedots, surfacedots, fr.x, &top_->atoms,
967 &top_->symtab, fr.ePBC, fr.box, bIncludeSolute_);
970 ah.startFrame(frnr, fr.time);
973 aah.startFrame(frnr, fr.time);
974 rah.startFrame(frnr, fr.time);
978 dgh.startFrame(frnr, fr.time);
981 ah.setPoint(0, totarea);
983 real totalArea, dgsolv;
984 if (bResAt || bDGsol)
986 computeAreas(surfaceSel, surfaceSel, frameData.atomAreas_, dgsFactor_,
987 &totalArea, &dgsolv, aah, rah, &frameData.res_a_);
990 dgh.setPoint(0, dgsolv);
993 for (size_t g = 0; g < outputSel.size(); ++g)
997 aah.selectDataSet(g + 1);
998 rah.selectDataSet(g + 1);
1000 computeAreas(surfaceSel, outputSel[g], frameData.atomAreas_, dgsFactor_,
1001 &totalArea, &dgsolv, aah, rah, &frameData.res_a_);
1002 ah.setPoint(g + 1, totalArea);
1005 dgh.setPoint(g + 1, dgsolv);
1023 for (int i = 0; i < surfaceSel.posCount(); ++i)
1025 totmass += surfaceSel.position(i).mass();
1027 const real density = totmass*AMU/(totvolume*NANO*NANO*NANO);
1028 vh.startFrame(frnr, fr.time);
1029 vh.setPoint(0, totvolume);
1030 vh.setPoint(1, density);
1036 Sasa::finishAnalysis(int /*nframes*/)
1040 // fp3 = ftp2FILE(efITP, nfile, fnm, "w");
1041 // fprintf(fp3, "[ position_restraints ]\n"
1042 // "#define FCX 1000\n"
1043 // "#define FCY 1000\n"
1044 // "#define FCZ 1000\n"
1045 // "; Atom Type fx fy fz\n");
1046 // for (i = 0; i < nx[0]; i++)
1048 // if (atom_area[i] > minarea)
1050 // fprintf(fp3, "%5d 1 FCX FCX FCZ\n", ii+1);
1066 const char SasaInfo::name[] = "sasa";
1067 const char SasaInfo::shortDescription[] =
1068 "Compute solvent accessible surface area";
1070 TrajectoryAnalysisModulePointer SasaInfo::create()
1072 return TrajectoryAnalysisModulePointer(new Sasa);
1075 } // namespace analysismodules