Add SAXS scoring module

examples/scoring-saxs/data/protein-protein_saxs.dat

@@ -0,0 +1,52 @@
+# Simulated scattering data for examples/docking-protein-protein/data/e2a-hpr_1GGR.pdb
+0.000000E+00 0.110408E+08
+0.100000E-01 0.109074E+08
+0.200000E-01 0.105167E+08
+0.300000E-01 0.989488E+07
+0.400000E-01 0.908313E+07
+0.500000E-01 0.813279E+07
+0.600000E-01 0.710047E+07
+0.700000E-01 0.604272E+07
+0.800000E-01 0.501118E+07
+0.900000E-01 0.404870E+07
+0.100000E+00 0.318682E+07
+0.110000E+00 0.244476E+07
+0.120000E+00 0.182985E+07
+0.130000E+00 0.133915E+07
+0.140000E+00 0.961881E+06
+0.150000E+00 0.682183E+06
+0.160000E+00 0.481886E+06
+0.170000E+00 0.342843E+06
+0.180000E+00 0.248682E+06
+0.190000E+00 0.185849E+06
+0.200000E+00 0.144003E+06
+0.210000E+00 0.115863E+06
+0.220000E+00 0.967074E+05
+0.230000E+00 0.836887E+05
+0.240000E+00 0.751560E+05
+0.250000E+00 0.700796E+05
+0.260000E+00 0.676557E+05
+0.270000E+00 0.670955E+05
+0.280000E+00 0.675737E+05
+0.290000E+00 0.682738E+05
+0.300000E+00 0.684887E+05
+0.310000E+00 0.677076E+05
+0.320000E+00 0.656805E+05
+0.330000E+00 0.624174E+05
+0.340000E+00 0.581470E+05
+0.350000E+00 0.532498E+05
+0.360000E+00 0.481699E+05
+0.370000E+00 0.433383E+05
+0.380000E+00 0.391132E+05
+0.390000E+00 0.357524E+05
+0.400000E+00 0.334000E+05
+0.410000E+00 0.320988E+05
+0.420000E+00 0.318072E+05
+0.430000E+00 0.324247E+05
+0.440000E+00 0.338126E+05
+0.450000E+00 0.358077E+05
+0.460000E+00 0.382327E+05
+0.470000E+00 0.409038E+05
+0.480000E+00 0.436358E+05
+0.490000E+00 0.462483E+05
+0.500000E+00 0.485731E+05

examples/scoring-saxs/saxsscore-test.cfg

@@ -0,0 +1,27 @@
+# ====================================================================
+# Scoring example
+
+# directory in which the scoring will be done
+run_dir = "saxsscore-test"
+ncores = 40
+mode = "local"
+
+# ensemble to be scored
+molecules = ["../scoring/data/protein-protein_1w.pdb",
+             "../scoring/data/protein-protein_2w.pdb"
+             ]
+
+
+# ====================================================================
+# Parameters for each stage are defined below
+
+[topoaa]
+autohis = true
+
+[emscoring]
+
+[saxsscore]
+saxs_fname = "data/protein-protein_saxs.dat"
+atsas_path = "/Applications/ATSAS"  # CHANGE IF INSTALLED ELSEWHERE
+
+# ====================================================================

integration_tests/test_saxs.py

@@ -0,0 +1,76 @@
+from haddock.modules.analysis.saxsscore import HaddockModule as SaxsModule
+
+import shutil
+from pathlib import Path
+import pytest
+from haddock.libs.libontology import PDBFile
+from tests import golden_data
+
+from haddock.modules.analysis.saxsscore import (
+    DEFAULT_CONFIG as DEFAULT_SAXSSCORE_CONFIG,
+)
+from haddock.modules.analysis.saxsscore.saxs import run_crysol
+
+
+class MockPreviousIO:
+    def __init__(self, path):
+        self.path = path
+
+    def retrieve_models(self, individualize: bool = False):
+        shutil.copy(
+            Path(golden_data, "e2aP_1F3G_haddock.pdb"),
+            Path(".", "e2aP_1F3G_haddock.pdb"),
+        )
+        model_list = [
+            PDBFile(
+                file_name="e2aP_1F3G_haddock.pdb",
+                path=".",
+                score=0.0,
+            ),
+        ]
+
+        return model_list
+
+    def output(self):
+        return None
+
+
+@pytest.fixture
+def saxsscore():
+    return SaxsModule(
+        order=0,
+        path=Path("."),
+        init_params=DEFAULT_SAXSSCORE_CONFIG,
+    )
+
+
+@pytest.fixture
+def saxs_data():
+    yield Path(golden_data, "saxs.dat")
+
+
+ATSAS_PATH = "/home/rodrigo/software/ATSAS-3.2.1-1"
+
+
+def test_saxsscore(saxsscore, saxs_data):
+
+    saxsscore.previous_io = MockPreviousIO(path=saxsscore.path)
+    saxsscore.params["saxs_fname"] = saxs_data
+    saxsscore.params["atsas_path"] = ATSAS_PATH
+
+    saxsscore.run()
+
+    assert Path("saxsscore.tsv").exists()
+
+
+def test_crysol_is_executable(saxs_data):
+    pdb_f = Path(golden_data, "hpr_ensemble_1_haddock.pdb")
+    run_crysol(
+        atsas_path=ATSAS_PATH,
+        input_f=pdb_f,
+        saxs_data=saxs_data,
+        lm=20,
+        ns=20,
+    )
+
+    assert Path(f"{Path(pdb_f.stem)}_saxs.fit").exists()

src/haddock/modules/analysis/saxsscore/__init__.py

@@ -0,0 +1,110 @@
+"""
+SAXS scoring module
+===================
+
+Requires ATSAS 2.6.0 or newer.
+
+This module adds a SAXS term to the HADDOCK score.
+The SAXS term is based on the fit (Chi^2) of a model to corresponding
+scattering data, as calculated by the ATSAS program CRYSOL.
+
+By default the weights of the HADDOCK score (calculated by the preceding module)
+and the Chi value are 1 and 50, respectively.
+If the preceding module did not provide HADDOCK scores,
+the Chi^2 values are returned as is (not multiplied by 50).
+
+The resulting scores are found in ``saxsscore.tsv``.
+This module does not alter the models themselves
+and passes them on unchanged to the subsequent module.
+
+For more information on SAXS-based scoring in HADDOCK, see
+
+Karaca E, Bonvin AMJJ. 2013. On the Usefulness of Ion-Mobility Mass Spectrometry
+and SAXS Data in Scoring Docking Decoys. *Acta Crystallographica* D69:683-694.
+"""
+
+import math
+from pathlib import Path
+
+from haddock.libs.libutil import check_subprocess
+from haddock.modules import BaseHaddockModule
+from haddock.modules.analysis.saxsscore.saxs import (
+    run_crysol,
+    read_chi2,
+    calculate_haddocksaxs_score,
+    output_saxsscore,
+)
+
+
+RECIPE_PATH = Path(__file__).resolve().parent
+DEFAULT_CONFIG = Path(RECIPE_PATH, "defaults.yaml")
+
+
+class HaddockModule(BaseHaddockModule):
+    """HADDOCK3 module to perform SAXS-based scoring."""
+
+    name = RECIPE_PATH.name
+
+    def __init__(self, order, path, *ignore, init_params=DEFAULT_CONFIG, 
+                 **everything):
+        super().__init__(order, path, init_params)
+
+    @classmethod
+    def confirm_installation(cls):
+        """Confirm this module is installed."""
+        check_subprocess("crysol -h")
+
+    def _run(self):
+        """Execute module."""
+        models_to_score = self.previous_io.retrieve_models(individualize=True)
+
+        # Get parameters from .yaml/.cfg
+        saxs_data = self.params["saxs_fname"]
+        lm = self.params["lm"]
+        ns = self.params["ns"]
+        cst = self.params["cst"]
+        w_haddock = self.params["w_haddock"]
+        w_saxs = self.params["w_saxs"]
+
+        for model in models_to_score:
+
+            # Run CRYSOL
+            run_crysol(
+                atsas_path=self.params["atsas_path"],
+                input_f=model.rel_path,
+                saxs_data=saxs_data,
+                lm=lm,
+                ns=ns,
+                cst=cst,
+            )
+
+            # Read Chi^2 value
+            fit_file = (Path(model.file_name).stem 
+                        + "_" + Path(saxs_data).stem + ".fit")
+            model.chi2 = read_chi2(fit_file=fit_file)
+
+            # Calculate HADDOCKsaxs score
+            if math.isnan(model.score):
+                haddocksaxs_score = model.chi2
+            else:
+                haddocksaxs_score = calculate_haddocksaxs_score(
+                    score=model.score,
+                    chi2=model.chi2,
+                    w_haddock=w_haddock,
+                    w_saxs=w_saxs,
+                )
+
+            model.score = haddocksaxs_score
+
+        # Write output
+        self.output_models = models_to_score
+        output_fname="saxsscore.tsv"
+        self.log(f"Saving output to {output_fname}")
+
+        output_saxsscore(
+            output_models=self.output_models,
+            output_fname=output_fname,
+        )
+
+        # Send models (unchanged) to the next step
+        self.export_io_models()

src/haddock/modules/analysis/saxsscore/defaults.yaml

@@ -0,0 +1,80 @@
+#
+# Group scoring
+#
+w_haddock:
+  default: 1.0
+  type: float
+  min: 0
+  max: 9999
+  precision: 3
+  title: Weight of the HADDOCK score
+  short: Weight of the HADDOCK score in HADDOCKsaxs.
+  long: Weight of the HADDOCK score in the HADDOCKsaxs score.
+    The HADDOCK score from the previous module is multiplied by this weight
+    and added to the chi term to get the HADDOCKsaxs score.
+  group: "scoring"
+  explevel: expert
+w_saxs:
+  default: 50.0
+  type: float
+  min: 0
+  max: 9999
+  precision: 3
+  title: Weight of the SAXS term
+  short: Weight of the SAXS term in the scoring function.
+  long: Weight of the SAXS term in the scoring function.
+    The chi value is multiplied by this weight
+    and added to the HADDOCK score to get the HADDOCKsaxs score.
+  group: "scoring"
+  explevel: expert
+#
+# Group CRYSOL
+#
+saxs_fname:
+  default: ""
+  type: file
+  title: Path to SAXS data
+  short: Path to SAXS data.
+  long: Path to the .dat file containing SAXS data for CRYSOL.
+    Two or three columns separated by whitespace, containing (in this order)
+    momentum transfer, scattering intensity, and (optional) standard deviation.
+  group: "scoring"
+  explevel: easy
+lm:
+  default: 18
+  type: integer
+  min: 1
+  max: 100
+  title: Maximum order of harmonics
+  short: Maximum order of harmonics.
+  long: Maximum order of harmonics for CRYSOL.
+  group: "scoring"
+  explevel: expert
+ns:
+  default: 256
+  type: integer
+  min: 1
+  max: 10001
+  title: Number of calculated data points in computed curve
+  short: Number of calculated data points in computed curve.
+  long: Number of calculated data points in computed curve from CRYSOL.
+  group: "scoring"
+  explevel: expert
+cst:
+  default: true
+  type: boolean
+  title: Subtract constant
+  short: Subtract constant.
+  long: Subtract constant in CRYSOL.
+  group: "scoring"
+  explevel: expert
+atsas_path:
+  default: "~/software/ATSAS"
+  type: string
+  minchars: 5
+  maxchars: 4095
+  title: Path to ATSAS installation
+  short: Path to ATSAS installation.
+  long: Path to the installation of ATSAS.
+  group: "scoring"
+  explevel: easy