Voroscoring

docs/INSTALL.md

@@ -154,3 +154,24 @@ on your machine.
 
 Please refer to the [official page](http://docs.openmm.org/latest/userguide/)
 of the project for a full description of the installation procedure.
+
+
+
+## `voroscoring`
+
+The use of the `[voroscoring]` module requires:
+- A cluster with SLURM installed
+- The setup of a conda environement (e.g.: ftdmp), in which you will install FTDMP
+- A functional installation of [FTDMP](https://github.com/kliment-olechnovic)
+
+Once those three conditions are fulfilled, when using the `[voroscoring]` module in haddock3, the configuration file must be tuned to contain parameters describing how to load the appropriate conda env (ftdmp) and where to find FTDMP scripts and executables:
+
+```TOML
+[voroscoring]
+# This parameter defines the base directory where conda / miniconda is installed
+conda_install_dir = "/absolute/path/to/conda/"
+# This parameter defines the name of the conda env that you created and where FTDMP is installled
+conda_env_name = "ftdmp"
+# This parameter defines where FTDMP scripts / executables can be found
+ftdmp_install_dir = "/absolute/path/to/FTDMP/"
+```

examples/scoring/voroscoring-test.cfg

@@ -0,0 +1,33 @@
+# ====================================================================
+# Scoring example
+
+# directory in which the scoring will be done
+run_dir = "run1-voroscoring-test"
+clean = false
+
+# execution mode
+ncores = 3
+mode = "local"
+
+# ensemble of different complexes to be scored
+molecules = ["data/T161-rescoring-ens.pdb",
+             "data/HY3.pdb",
+             "data/protein-dna_1w.pdb",
+             "data/protein-protein_1w.pdb", 
+             "data/protein-protein_2w.pdb",
+             "data/protein-trimer_1w.pdb"
+             ]
+
+# ====================================================================
+# Parameters for each stage are defined below
+
+[topoaa]
+
+[voroscoring]
+
+[seletop]
+select = 3
+
+[caprieval]
+
+# ====================================================================

src/haddock/modules/scoring/__init__.py

@@ -1,7 +1,8 @@
 """HADDOCK3 modules to score models."""
+from os import linesep
 import pandas as pd
 
-from haddock.core.typing import FilePath, Path, Any
+from haddock.core.typing import FilePath, Path, Any, Optional
 from haddock.modules.base_cns_module import BaseCNSModule
 from haddock.modules import BaseHaddockModule, PDBFile
 
@@ -14,6 +15,7 @@ def output(
             output_fname: FilePath,
             sep: str = "\t",
             ascending_sort: bool = True,
+            header_comments: Optional[str] = None,
             ) -> None:
         r"""Save the output in comprehensive tables.
 
@@ -36,11 +38,23 @@ def output(
         df_sc = pd.DataFrame(sc_data, columns=df_columns)
         df_sc_sorted = df_sc.sort_values(by="score", ascending=ascending_sort)
         # writes to disk
-        df_sc_sorted.to_csv(output_fname,
-                            sep=sep,
-                            index=False,
-                            na_rep="None",
-                            float_format="%.3f")
+        output_file = open(output_fname, 'a')
+        # Check if some comment in header are here
+        if header_comments:
+            # Make sure the comments is ending by a new line
+            if header_comments[-1] != linesep:
+                header_comments += linesep
+            output_file.write(header_comments)
+        # Write the dataframe
+        df_sc_sorted.to_csv(
+            output_file,
+            sep=sep,
+            index=False,
+            na_rep="None",
+            float_format="%.3f",
+            lineterminator=linesep,
+            )
+
         return
 
 

src/haddock/modules/scoring/emscoring/__init__.py

@@ -1,7 +1,8 @@
 """EM scoring module.
 
-This module performs energy minimization and scoring of the models generated in
-the previous step of the workflow. No restraints are applied during this step.
+This module performs energy minimization and scoring of the models generated
+in the previous step of the workflow.
+Note that no restraints (AIRs) are applied during this step.
 """
 
 from pathlib import Path

src/haddock/modules/scoring/mdscoring/__init__.py

@@ -1,7 +1,8 @@
 """MD scoring module.
 
 This module will perform a short MD simulation on the input models and
-score them. No restraints are applied during this step.
+score them.
+Note that no restraints (AIRs) are applied during this step.
 """
 
 from pathlib import Path

src/haddock/modules/scoring/voroscoring/__init__.py

@@ -0,0 +1,94 @@
+"""Voro scoring module.
+
+This module performs scoring of input pdb models using ftdmp voro-mqa-all tool.
+For more information, please check: https://github.com/kliment-olechnovic/ftdmp
+
+It is a third party module, and requires the appropriate set up and intallation
+for it to run without issue.
+"""
+from os import linesep
+from pathlib import Path
+
+from haddock.core.defaults import MODULE_DEFAULT_YAML
+from haddock.core.typing import Any, FilePath
+from haddock.modules import get_engine
+from haddock.modules.scoring import ScoringModule
+from haddock.modules.scoring.voroscoring.voroscoring import (
+    VoroMQA,
+    update_models_with_scores,
+    )
+
+RECIPE_PATH = Path(__file__).resolve().parent
+DEFAULT_CONFIG = Path(RECIPE_PATH, MODULE_DEFAULT_YAML)
+
+
+class HaddockModule(ScoringModule):
+    """."""
+
+    name = RECIPE_PATH.name
+
+    def __init__(
+            self,
+            order: int,
+            path: Path,
+            *ignore: Any,
+            init_params: FilePath = DEFAULT_CONFIG,
+            **everything: Any,
+            ) -> None:
+        """Initialize class."""
+        super().__init__(order, path, init_params)
+
+    @classmethod
+    def confirm_installation(cls) -> None:
+        """Confirm module is installed."""
+        # FIXME ? Check :
+        # - if conda env is accessible
+        # - if ftdmp is accessible
+        return
+
+    def _run(self) -> None:
+        """Execute module."""
+        # Retrieve previous models
+        try:
+            models_to_score = self.previous_io.retrieve_models(
+                individualize=True
+                )
+        except Exception as e:
+            self.finish_with_error(e)
+
+        # Initiate VoroMQA object
+        output_fname = Path("voro_mqa_all.tsv")
+        voromqa = VoroMQA(
+            models_to_score,
+            './',
+            self.params,
+            output=output_fname,
+            )
+
+        # Launch machinery
+        jobs: list[VoroMQA] = [voromqa]
+        # Run Job(s)
+        self.log("Running Voro-mqa scoring")
+        Engine = get_engine(self.params['mode'], self.params)
+        engine = Engine(jobs)
+        engine.run()
+        self.log("Voro-mqa scoring finished!")
+
+        # Update score of output models
+        try:
+            self.output_models = update_models_with_scores(
+                output_fname,
+                models_to_score,
+                metric=self.params["metric"],
+                )
+        except ValueError as e:
+            self.finish_with_error(e)
+
+        # Write output file
+        scoring_tsv_fpath = f"{RECIPE_PATH.name}.tsv"
+        self.output(
+            scoring_tsv_fpath,
+            header_comments=f"# Note that negative of the value are reported in the case of non-energetical predictions{linesep}",  # noqa : E501
+            )
+        # Export to next module
+        self.export_io_models()

src/haddock/modules/scoring/voroscoring/defaults.yaml

@@ -0,0 +1,77 @@
+metric:
+  default: jury_score
+  type: string
+  choices:
+    - jury_score
+    - GNN_sum_score
+    - GNN_pcadscore
+    - voromqa_dark
+    - voromqa_light
+    - voromqa_energy
+    - gen_voromqa_energy
+    - clash_score
+    - area
+  minchars: 1
+  maxchars: 50
+  title: VoroMQA metric used to score.
+  short: VoroMQA metric used to score.
+  long: VoroMQA metric used to score.
+  group: analysis
+  explevel: easy
+
+conda_install_dir:
+  default: "/trinity/login/vreys/miniconda3/"
+  type: string
+  minchars: 1
+  maxchars: 158
+  title: Path to conda intall directory.
+  short: Absolute path to conda intall directory.
+  long: Absolute path to conda intall directory.
+  group: execution
+  explevel: easy
+
+conda_env_name:
+  default: "ftdmp5"
+  type: string
+  minchars: 1
+  maxchars: 100
+  title: Name of the ftdmp conda env.
+  short: Name of the ftdmp conda env.
+  long: Name of the ftdmp conda env.
+  group: execution
+  explevel: easy
+
+ftdmp_install_dir:
+  default: "/trinity/login/vreys/Venclovas/ftdmp/"
+  type: string
+  minchars: 1
+  maxchars: 158
+  title: Path to ftdmp intall directory.
+  short: Absolute path to ftdmp intall directory.
+  long: Absolute path to ftdmp intall directory.
+  group: execution
+  explevel: easy
+
+nb_gpus:
+  default: 1
+  type: integer
+  min: 1
+  max: 420
+  title: Number of accessible gpu on the device.
+  short: Number of accessible gpu on the device.
+  long: Number of accessible gpu on the device.
+  group: execution
+  explevel: easy
+
+concat_chain_:
+  default: []
+  type: list
+  minitems: 0
+  maxitems: 100
+  title: List of residues supposed to be buried
+  short: List of residues supposed to be buried
+  long: concat_chain_* is an expandable parameter. You can provide concat_chain_1,
+    concat_chain_2, concat_chain_3, etc. For each selection, enlisted chains will
+    be concatenated as one prior to scoring.
+  group: analysis
+  explevel: expert