From b25e2de4ded679e911c2b2d1fb9ad34c86662734 Mon Sep 17 00:00:00 2001
From: Domantas Kuryla <dk584@icepc16>
Date: Sat, 20 Dec 2025 20:38:37 +0000
Subject: [PATCH 1/9] Add NCIA R739x5 calculation script

---
 .../ncia_r739x5/.dvc/.gitignore               |   3 +
 .../ncia_r739x5/.dvc/config                   |   0
 .../ncia_r739x5/.dvcignore                    |   3 +
 .../ncia_r739x5/calc_ncia_r739x5.py           | 167 ++++++++++++++++++
 4 files changed, 173 insertions(+)
 create mode 100644 ml_peg/calcs/non_covalent_interactions/ncia_r739x5/.dvc/.gitignore
 create mode 100644 ml_peg/calcs/non_covalent_interactions/ncia_r739x5/.dvc/config
 create mode 100644 ml_peg/calcs/non_covalent_interactions/ncia_r739x5/.dvcignore
 create mode 100644 ml_peg/calcs/non_covalent_interactions/ncia_r739x5/calc_ncia_r739x5.py

diff --git a/ml_peg/calcs/non_covalent_interactions/ncia_r739x5/.dvc/.gitignore b/ml_peg/calcs/non_covalent_interactions/ncia_r739x5/.dvc/.gitignore
new file mode 100644
index 000000000..528f30c71
--- /dev/null
+++ b/ml_peg/calcs/non_covalent_interactions/ncia_r739x5/.dvc/.gitignore
@@ -0,0 +1,3 @@
+/config.local
+/tmp
+/cache
diff --git a/ml_peg/calcs/non_covalent_interactions/ncia_r739x5/.dvc/config b/ml_peg/calcs/non_covalent_interactions/ncia_r739x5/.dvc/config
new file mode 100644
index 000000000..e69de29bb
diff --git a/ml_peg/calcs/non_covalent_interactions/ncia_r739x5/.dvcignore b/ml_peg/calcs/non_covalent_interactions/ncia_r739x5/.dvcignore
new file mode 100644
index 000000000..519730552
--- /dev/null
+++ b/ml_peg/calcs/non_covalent_interactions/ncia_r739x5/.dvcignore
@@ -0,0 +1,3 @@
+# Add patterns of files dvc should ignore, which could improve
+# the performance. Learn more at
+# https://dvc.org/doc/user-guide/dvcignore
diff --git a/ml_peg/calcs/non_covalent_interactions/ncia_r739x5/calc_ncia_r739x5.py b/ml_peg/calcs/non_covalent_interactions/ncia_r739x5/calc_ncia_r739x5.py
new file mode 100644
index 000000000..e68ea7d87
--- /dev/null
+++ b/ml_peg/calcs/non_covalent_interactions/ncia_r739x5/calc_ncia_r739x5.py
@@ -0,0 +1,167 @@
+"""
+Compute the NCIA R739x5 repulsive contacts dataset.
+
+Journal of Chemical Theory and Computation 2021 17 (3), 1548-1561.
+"""
+
+from __future__ import annotations
+
+from pathlib import Path
+
+from ase import units
+from ase.io import read, write
+import mlipx
+from mlipx.abc import NodeWithCalculator
+import numpy as np
+from tqdm import tqdm
+import zntrack
+
+from ml_peg.calcs.utils.utils import chdir, download_s3_data
+from ml_peg.models.get_models import load_models
+from ml_peg.models.models import current_models
+
+MODELS = load_models(current_models)
+
+KCAL_TO_EV = units.kcal / units.mol
+EV_TO_KCAL = 1 / KCAL_TO_EV
+
+OUT_PATH = Path(__file__).parent / "outputs"
+EXCLUDE_NOBLE_GASES = True
+
+
+class NCIAR739x5Benchmark(zntrack.Node):
+    """Benchmark the NCIA_R739x5 repulsive contacts dataset."""
+
+    model: NodeWithCalculator = zntrack.deps()
+    model_name: str = zntrack.params()
+
+    def get_ref_energies(self, data_path):
+        """
+        Get reference energies.
+
+        Parameters
+        ----------
+        data_path
+            Path to data.
+        """
+        self.ref_energies = {}
+        with open(data_path / "NCIA_R739x5_benchmark.txt") as lines:
+            for i, line in enumerate(lines):
+                if i < 3:
+                    continue
+                items = line.strip().split()
+                label = items[0]
+                ref_energy = float(items[1]) * KCAL_TO_EV
+                self.ref_energies[label] = ref_energy
+
+    @staticmethod
+    def get_monomers(atoms):
+        """
+        Get ASE atoms objects of the monomers.
+
+        Parameters
+        ----------
+        atoms
+            ASE atoms object of the structure.
+
+        Returns
+        -------
+        tuple[ASE.Atoms, ASE.Atoms]
+            Tuple containing the two monomers.
+        """
+        if isinstance(atoms.info["selection_a"], str):
+            a_ids = [int(id) for id in atoms.info["selection_a"].split("-")]
+            a_ids[0] -= 1
+        else:
+            a_ids = [int(atoms.info["selection_a"]) - 1, int(atoms.info["selection_a"])]
+
+        if isinstance(atoms.info["selection_b"], str):
+            b_ids = [int(id) for id in atoms.info["selection_b"].split("-")]
+            b_ids[0] -= 1
+        else:
+            b_ids = [int(atoms.info["selection_b"]) - 1, int(atoms.info["selection_b"])]
+
+        atoms_a = atoms[a_ids[0] : a_ids[1]]
+        atoms_b = atoms[b_ids[0] : b_ids[1]]
+        assert len(atoms_a) + len(atoms_b) == len(atoms)
+
+        atoms_a.info["charge"] = int(atoms.info["charge_a"])
+        atoms_a.info["spin"] = 1
+
+        atoms_b.info["charge"] = int(atoms.info["charge_b"])
+        atoms_b.info["spin"] = 1
+        return (atoms_a, atoms_b)
+
+    def run(self):
+        """Run new benchmark."""
+        # Read in data and attach calculator
+        data_path = (
+            download_s3_data(
+                filename="NCIA_R739.zip",
+                key="inputs/non_covalent_interactions/NCIA_R739x5/NCIA_R739x5.zip",
+            )
+            / "NCIA_739x5"
+        )
+        self.get_ref_energies(data_path)
+
+        calc = self.model.get_calculator()
+
+        for label, ref_energy in tqdm(self.ref_energies.items()):
+            xyz_fname = f"{label}.xyz"
+            atoms = read(data_path / "geometries" / xyz_fname)
+            # exclude noble gases
+            if (
+                np.isin(atoms.numbers, [2, 10, 18, 36, 54, 86]).any()
+                and EXCLUDE_NOBLE_GASES
+            ):
+                continue
+            atoms_a, atoms_b = self.get_monomers(atoms)
+            atoms.info["spin"] = 1
+            atoms.info["charge"] = int(atoms_a.info["charge"] + atoms_b.info["charge"])
+            atoms.calc = calc
+            atoms_a.calc = calc
+            atoms_b.calc = calc
+
+            atoms.info["model_int_energy"] = (
+                atoms.get_potential_energy()
+                - atoms_a.get_potential_energy()
+                - atoms_b.get_potential_energy()
+            )
+            atoms.info["ref_int_energy"] = ref_energy
+            atoms.calc = None
+
+            write_dir = OUT_PATH / self.model_name
+            write_dir.mkdir(parents=True, exist_ok=True)
+            write(write_dir / f"{label}.xyz", atoms)
+
+
+def build_project(repro: bool = False) -> None:
+    """
+    Build mlipx project.
+
+    Parameters
+    ----------
+    repro
+        Whether to call dvc repro -f after building.
+    """
+    project = mlipx.Project()
+    benchmark_node_dict = {}
+
+    for model_name, model in MODELS.items():
+        with project.group(model_name):
+            benchmark = NCIAR739x5Benchmark(
+                model=model,
+                model_name=model_name,
+            )
+            benchmark_node_dict[model_name] = benchmark
+
+    if repro:
+        with chdir(Path(__file__).parent):
+            project.repro(build=True, force=True)
+    else:
+        project.build()
+
+
+def test_ncia_r739x5():
+    """Run NCIA_R739x5 barriers benchmark via Pytest."""
+    build_project(repro=True)

From f80048afaa02c28b099bce51d38ce82a45166b23 Mon Sep 17 00:00:00 2001
From: Domantas Kuryla <dk584@icepc16>
Date: Sun, 21 Dec 2025 21:13:00 +0000
Subject: [PATCH 2/9] Add NCIA_R739x5 analysis files

---
 .../ncia_r739x5/analyse_ncia_r739x5.py        | 177 ++++++++++++++++++
 .../ncia_r739x5/metrics.yml                   |  13 ++
 2 files changed, 190 insertions(+)
 create mode 100644 ml_peg/analysis/non_covalent_interactions/ncia_r739x5/analyse_ncia_r739x5.py
 create mode 100644 ml_peg/analysis/non_covalent_interactions/ncia_r739x5/metrics.yml

diff --git a/ml_peg/analysis/non_covalent_interactions/ncia_r739x5/analyse_ncia_r739x5.py b/ml_peg/analysis/non_covalent_interactions/ncia_r739x5/analyse_ncia_r739x5.py
new file mode 100644
index 000000000..ba30fc7d2
--- /dev/null
+++ b/ml_peg/analysis/non_covalent_interactions/ncia_r739x5/analyse_ncia_r739x5.py
@@ -0,0 +1,177 @@
+"""Analyse ncia_r739x5 benchmark."""
+
+from __future__ import annotations
+
+from pathlib import Path
+
+from ase import units
+from ase.io import read, write
+import pytest
+
+from ml_peg.analysis.utils.decorators import build_table, plot_parity
+from ml_peg.analysis.utils.utils import (
+    build_d3_name_map,
+    load_metrics_config,
+    mae,
+    rmse,
+)
+from ml_peg.app import APP_ROOT
+from ml_peg.calcs import CALCS_ROOT
+from ml_peg.models.get_models import load_models
+from ml_peg.models.models import current_models
+
+MODELS = load_models(current_models)
+D3_MODEL_NAMES = build_d3_name_map(MODELS)
+
+KCAL_TO_EV = units.kcal / units.mol
+EV_TO_KCAL = 1 / KCAL_TO_EV
+CALC_PATH = CALCS_ROOT / "non_covalent_interactions" / "ncia_r739x5" / "outputs"
+OUT_PATH = APP_ROOT / "data" / "non_covalent_interactions" / "ncia_r739x5"
+
+METRICS_CONFIG_PATH = Path(__file__).with_name("metrics.yml")
+DEFAULT_THRESHOLDS, DEFAULT_TOOLTIPS, DEFAULT_WEIGHTS = load_metrics_config(
+    METRICS_CONFIG_PATH
+)
+
+
+def labels() -> list:
+    """
+    Get list of system names.
+
+    Returns
+    -------
+    list
+        List of all system names.
+    """
+    for model in MODELS:
+        labels_list = [path.stem for path in (CALC_PATH / model).glob("*.xyz")]
+        break
+    return labels_list
+
+
+@pytest.fixture
+@plot_parity(
+    filename=OUT_PATH / "figure_ncia_r739x5.json",
+    title="Interaction energies",
+    x_label="Predicted energy / eV",
+    y_label="Reference energy / eV",
+    hoverdata={
+        "Labels": labels(),
+    },
+)
+def interaction_energies() -> dict[str, list]:
+    """
+    Get interaction energies for all systems.
+
+    Returns
+    -------
+    dict[str, list]
+        Dictionary of all reference and predicted interaction energies.
+    """
+    results = {"ref": []} | {mlip: [] for mlip in MODELS}
+
+    ref_stored = False
+
+    for model_name in MODELS:
+        for label in labels():
+            atoms = read(CALC_PATH / model_name / f"{label}.xyz")
+            if not ref_stored:
+                results["ref"].append(atoms.info["ref_int_energy"])
+
+            results[model_name].append(atoms.info["model_int_energy"])
+
+            # Write structures for app
+            structs_dir = OUT_PATH / model_name
+            structs_dir.mkdir(parents=True, exist_ok=True)
+            write(structs_dir / f"{label}.xyz", atoms)
+
+        ref_stored = True
+    return results
+
+
+@pytest.fixture
+def get_mae(interaction_energies) -> dict[str, float]:
+    """
+    Get mean absolute error for energies.
+
+    Parameters
+    ----------
+    interaction_energies
+        Dictionary of reference and predicted energies.
+
+    Returns
+    -------
+    dict[str, float]
+        Dictionary of predicted energy errors for all models.
+    """
+    results = {}
+    for model_name in MODELS:
+        results[model_name] = mae(
+            interaction_energies["ref"], interaction_energies[model_name]
+        )
+    return results
+
+
+@pytest.fixture
+def get_rmse(interaction_energies) -> dict[str, float]:
+    """
+    Get root mean square error for energies.
+
+    Parameters
+    ----------
+    interaction_energies
+        Dictionary of reference and predicted energies.
+
+    Returns
+    -------
+    dict[str, float]
+        Dictionary of predicted energy errors for all models.
+    """
+    results = {}
+    for model_name in MODELS:
+        results[model_name] = rmse(
+            interaction_energies["ref"], interaction_energies[model_name]
+        )
+    return results
+
+
+@pytest.fixture
+@build_table(
+    filename=OUT_PATH / "ncia_r739x5_metrics_table.json",
+    metric_tooltips=DEFAULT_TOOLTIPS,
+    thresholds=DEFAULT_THRESHOLDS,
+    mlip_name_map=D3_MODEL_NAMES,
+)
+def metrics(get_mae: dict[str, float], get_rmse: dict[str, float]) -> dict[str, dict]:
+    """
+    Get all metrics.
+
+    Parameters
+    ----------
+    get_mae
+        Mean absolute errors for all models.
+
+    get_rmse
+        Root Mean Square Error for all models.
+
+    Returns
+    -------
+    dict[str, dict]
+        Metric names and values for all models.
+    """
+    return {
+        "MAE": get_mae,
+        "RMSE": get_rmse,
+    }
+
+
+def test_ncia_r739x5(metrics: dict[str, dict]) -> None:
+    """
+    Run ncia_r739x5 test.
+
+    Parameters
+    ----------
+    metrics
+        All new benchmark metric names and dictionary of values for each model.
+    """
+    return
diff --git a/ml_peg/analysis/non_covalent_interactions/ncia_r739x5/metrics.yml b/ml_peg/analysis/non_covalent_interactions/ncia_r739x5/metrics.yml
new file mode 100644
index 000000000..cbfb7e7a2
--- /dev/null
+++ b/ml_peg/analysis/non_covalent_interactions/ncia_r739x5/metrics.yml
@@ -0,0 +1,13 @@
+metrics:
+  MAE:
+    good: 0.0
+    bad: 0.5
+    unit: eV
+    tooltip: Mean Absolute Error for all systems
+    level_of_theory: CCSD(T)
+  RMSE:
+    good: 0.0
+    bad: 0.5
+    unit: eV
+    tooltip: Root Mean Square Error for all systems
+    level_of_theory: CCSD(T)

From 6cf076d47a51efab17c017a8bb844be23283c681 Mon Sep 17 00:00:00 2001
From: Domantas Kuryla <dk584@icepc16>
Date: Mon, 22 Dec 2025 13:48:30 +0000
Subject: [PATCH 3/9] Fix directory name

---
 .../non_covalent_interactions/ncia_r739x5/calc_ncia_r739x5.py   | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/ml_peg/calcs/non_covalent_interactions/ncia_r739x5/calc_ncia_r739x5.py b/ml_peg/calcs/non_covalent_interactions/ncia_r739x5/calc_ncia_r739x5.py
index e68ea7d87..b85548b9a 100644
--- a/ml_peg/calcs/non_covalent_interactions/ncia_r739x5/calc_ncia_r739x5.py
+++ b/ml_peg/calcs/non_covalent_interactions/ncia_r739x5/calc_ncia_r739x5.py
@@ -100,7 +100,7 @@ def run(self):
                 filename="NCIA_R739.zip",
                 key="inputs/non_covalent_interactions/NCIA_R739x5/NCIA_R739x5.zip",
             )
-            / "NCIA_739x5"
+            / "NCIA_R739x5"
         )
         self.get_ref_energies(data_path)
 

From 26d7f20b158ed52aad4d24374c9b523bb45ae3f5 Mon Sep 17 00:00:00 2001
From: Domantas Kuryla <dk584@icepc16>
Date: Mon, 29 Dec 2025 20:20:17 +0000
Subject: [PATCH 4/9] Add D3 calc

---
 .../non_covalent_interactions/ncia_r739x5/calc_ncia_r739x5.py   | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/ml_peg/calcs/non_covalent_interactions/ncia_r739x5/calc_ncia_r739x5.py b/ml_peg/calcs/non_covalent_interactions/ncia_r739x5/calc_ncia_r739x5.py
index b85548b9a..77b1a7f1a 100644
--- a/ml_peg/calcs/non_covalent_interactions/ncia_r739x5/calc_ncia_r739x5.py
+++ b/ml_peg/calcs/non_covalent_interactions/ncia_r739x5/calc_ncia_r739x5.py
@@ -105,6 +105,8 @@ def run(self):
         self.get_ref_energies(data_path)
 
         calc = self.model.get_calculator()
+        # Add D3 calculator for this test
+        calc = self.model.add_d3_calculator(calc)
 
         for label, ref_energy in tqdm(self.ref_energies.items()):
             xyz_fname = f"{label}.xyz"

From f4b8efc211835f49be5fe8ce936d8790963d7aab Mon Sep 17 00:00:00 2001
From: joehart2001 <jh2536@cam.ac.uk>
Date: Wed, 7 Jan 2026 14:09:00 +0000
Subject: [PATCH 5/9] ncia_r739x5 app

---
 .../ncia_r739x5/analyse_ncia_r739x5.py        | 38 +-------
 .../ncia_r739x5/metrics.yml                   |  6 --
 .../ncia_r739x5/app_ncia_r739x5.py            | 91 +++++++++++++++++++
 3 files changed, 94 insertions(+), 41 deletions(-)
 create mode 100644 ml_peg/app/non_covalent_interactions/ncia_r739x5/app_ncia_r739x5.py

diff --git a/ml_peg/analysis/non_covalent_interactions/ncia_r739x5/analyse_ncia_r739x5.py b/ml_peg/analysis/non_covalent_interactions/ncia_r739x5/analyse_ncia_r739x5.py
index ba30fc7d2..7ebc5e467 100644
--- a/ml_peg/analysis/non_covalent_interactions/ncia_r739x5/analyse_ncia_r739x5.py
+++ b/ml_peg/analysis/non_covalent_interactions/ncia_r739x5/analyse_ncia_r739x5.py
@@ -9,12 +9,7 @@
 import pytest
 
 from ml_peg.analysis.utils.decorators import build_table, plot_parity
-from ml_peg.analysis.utils.utils import (
-    build_d3_name_map,
-    load_metrics_config,
-    mae,
-    rmse,
-)
+from ml_peg.analysis.utils.utils import build_d3_name_map, load_metrics_config, mae
 from ml_peg.app import APP_ROOT
 from ml_peg.calcs import CALCS_ROOT
 from ml_peg.models.get_models import load_models
@@ -44,7 +39,7 @@ def labels() -> list:
         List of all system names.
     """
     for model in MODELS:
-        labels_list = [path.stem for path in (CALC_PATH / model).glob("*.xyz")]
+        labels_list = sorted([path.stem for path in (CALC_PATH / model).glob("*.xyz")])
         break
     return labels_list
 
@@ -112,29 +107,6 @@ def get_mae(interaction_energies) -> dict[str, float]:
     return results
 
 
-@pytest.fixture
-def get_rmse(interaction_energies) -> dict[str, float]:
-    """
-    Get root mean square error for energies.
-
-    Parameters
-    ----------
-    interaction_energies
-        Dictionary of reference and predicted energies.
-
-    Returns
-    -------
-    dict[str, float]
-        Dictionary of predicted energy errors for all models.
-    """
-    results = {}
-    for model_name in MODELS:
-        results[model_name] = rmse(
-            interaction_energies["ref"], interaction_energies[model_name]
-        )
-    return results
-
-
 @pytest.fixture
 @build_table(
     filename=OUT_PATH / "ncia_r739x5_metrics_table.json",
@@ -142,7 +114,7 @@ def get_rmse(interaction_energies) -> dict[str, float]:
     thresholds=DEFAULT_THRESHOLDS,
     mlip_name_map=D3_MODEL_NAMES,
 )
-def metrics(get_mae: dict[str, float], get_rmse: dict[str, float]) -> dict[str, dict]:
+def metrics(get_mae: dict[str, float]) -> dict[str, dict]:
     """
     Get all metrics.
 
@@ -151,9 +123,6 @@ def metrics(get_mae: dict[str, float], get_rmse: dict[str, float]) -> dict[str,
     get_mae
         Mean absolute errors for all models.
 
-    get_rmse
-        Root Mean Square Error for all models.
-
     Returns
     -------
     dict[str, dict]
@@ -161,7 +130,6 @@ def metrics(get_mae: dict[str, float], get_rmse: dict[str, float]) -> dict[str,
     """
     return {
         "MAE": get_mae,
-        "RMSE": get_rmse,
     }
 
 
diff --git a/ml_peg/analysis/non_covalent_interactions/ncia_r739x5/metrics.yml b/ml_peg/analysis/non_covalent_interactions/ncia_r739x5/metrics.yml
index cbfb7e7a2..8118e8238 100644
--- a/ml_peg/analysis/non_covalent_interactions/ncia_r739x5/metrics.yml
+++ b/ml_peg/analysis/non_covalent_interactions/ncia_r739x5/metrics.yml
@@ -5,9 +5,3 @@ metrics:
     unit: eV
     tooltip: Mean Absolute Error for all systems
     level_of_theory: CCSD(T)
-  RMSE:
-    good: 0.0
-    bad: 0.5
-    unit: eV
-    tooltip: Root Mean Square Error for all systems
-    level_of_theory: CCSD(T)
diff --git a/ml_peg/app/non_covalent_interactions/ncia_r739x5/app_ncia_r739x5.py b/ml_peg/app/non_covalent_interactions/ncia_r739x5/app_ncia_r739x5.py
new file mode 100644
index 000000000..faee0c4dd
--- /dev/null
+++ b/ml_peg/app/non_covalent_interactions/ncia_r739x5/app_ncia_r739x5.py
@@ -0,0 +1,91 @@
+"""Run NCIA_R739x5 app."""
+
+from __future__ import annotations
+
+from dash import Dash
+from dash.html import Div
+
+from ml_peg.app import APP_ROOT
+from ml_peg.app.base_app import BaseApp
+from ml_peg.app.utils.build_callbacks import (
+    plot_from_table_column,
+    struct_from_scatter,
+)
+from ml_peg.app.utils.load import read_plot
+from ml_peg.models.get_models import get_model_names
+from ml_peg.models.models import current_models
+
+MODELS = get_model_names(current_models)
+BENCHMARK_NAME = "NCIA_R739x5"
+DOCS_URL = (
+    "https://ddmms.github.io/ml-peg/user_guide/benchmarks/"
+    "non_covalent_interactions.html#ncia-r739x5"
+)
+DATA_PATH = APP_ROOT / "data" / "non_covalent_interactions" / "ncia_r739x5"
+
+
+class NCIAR739x5App(BaseApp):
+    """NCIA_R739x5 benchmark app layout and callbacks."""
+
+    def register_callbacks(self) -> None:
+        """Register callbacks to app."""
+        scatter = read_plot(
+            DATA_PATH / "figure_ncia_r739x5.json",
+            id=f"{BENCHMARK_NAME}-figure",
+        )
+
+        model_dir = DATA_PATH / MODELS[0]
+        if model_dir.exists():
+            labels = sorted([f.stem for f in model_dir.glob("*.xyz")])
+            structs = [
+                f"assets/non_covalent_interactions/ncia_r739x5/{MODELS[0]}/{label}.xyz"
+                for label in labels
+            ]
+        else:
+            structs = []
+
+        plot_from_table_column(
+            table_id=self.table_id,
+            plot_id=f"{BENCHMARK_NAME}-figure-placeholder",
+            column_to_plot={"MAE": scatter},
+        )
+
+        struct_from_scatter(
+            scatter_id=f"{BENCHMARK_NAME}-figure",
+            struct_id=f"{BENCHMARK_NAME}-struct-placeholder",
+            structs=structs,
+            mode="struct",
+        )
+
+
+def get_app() -> NCIAR739x5App:
+    """
+    Get NCIA_R739x5 benchmark app layout and callback registration.
+
+    Returns
+    -------
+    NCIAR739x5App
+        Benchmark layout and callback registration.
+    """
+    return NCIAR739x5App(
+        name=BENCHMARK_NAME,
+        description=(
+            "Performance in predicting radical interaction energies "
+            "for the NCIA R739x5 dataset (radical complexes). "
+            "Reference data from CCSD(T) calculations."
+        ),
+        docs_url=DOCS_URL,
+        table_path=DATA_PATH / "ncia_r739x5_metrics_table.json",
+        extra_components=[
+            Div(id=f"{BENCHMARK_NAME}-figure-placeholder"),
+            Div(id=f"{BENCHMARK_NAME}-struct-placeholder"),
+        ],
+    )
+
+
+if __name__ == "__main__":
+    full_app = Dash(__name__, assets_folder=DATA_PATH.parent.parent)
+    benchmark_app = get_app()
+    full_app.layout = benchmark_app.layout
+    benchmark_app.register_callbacks()
+    full_app.run(port=8060, debug=True)

From 54ea0fff53eb8d7a9497eb0788dd93092b8949cc Mon Sep 17 00:00:00 2001
From: ElliottKasoar <45317199+ElliottKasoar@users.noreply.github.com>
Date: Fri, 30 Jan 2026 15:13:22 +0000
Subject: [PATCH 6/9] Remove mlipx

---
 .../ncia_r739x5/calc_ncia_r739x5.py           | 248 ++++++++----------
 1 file changed, 114 insertions(+), 134 deletions(-)

diff --git a/ml_peg/calcs/non_covalent_interactions/ncia_r739x5/calc_ncia_r739x5.py b/ml_peg/calcs/non_covalent_interactions/ncia_r739x5/calc_ncia_r739x5.py
index 77b1a7f1a..18d02e675 100644
--- a/ml_peg/calcs/non_covalent_interactions/ncia_r739x5/calc_ncia_r739x5.py
+++ b/ml_peg/calcs/non_covalent_interactions/ncia_r739x5/calc_ncia_r739x5.py
@@ -7,16 +7,15 @@
 from __future__ import annotations
 
 from pathlib import Path
+from typing import Any
 
-from ase import units
+from ase import Atoms, units
 from ase.io import read, write
-import mlipx
-from mlipx.abc import NodeWithCalculator
 import numpy as np
+import pytest
 from tqdm import tqdm
-import zntrack
 
-from ml_peg.calcs.utils.utils import chdir, download_s3_data
+from ml_peg.calcs.utils.utils import download_s3_data
 from ml_peg.models.get_models import load_models
 from ml_peg.models.models import current_models
 
@@ -29,141 +28,122 @@
 EXCLUDE_NOBLE_GASES = True
 
 
-class NCIAR739x5Benchmark(zntrack.Node):
-    """Benchmark the NCIA_R739x5 repulsive contacts dataset."""
-
-    model: NodeWithCalculator = zntrack.deps()
-    model_name: str = zntrack.params()
-
-    def get_ref_energies(self, data_path):
-        """
-        Get reference energies.
-
-        Parameters
-        ----------
-        data_path
-            Path to data.
-        """
-        self.ref_energies = {}
-        with open(data_path / "NCIA_R739x5_benchmark.txt") as lines:
-            for i, line in enumerate(lines):
-                if i < 3:
-                    continue
-                items = line.strip().split()
-                label = items[0]
-                ref_energy = float(items[1]) * KCAL_TO_EV
-                self.ref_energies[label] = ref_energy
-
-    @staticmethod
-    def get_monomers(atoms):
-        """
-        Get ASE atoms objects of the monomers.
-
-        Parameters
-        ----------
-        atoms
-            ASE atoms object of the structure.
-
-        Returns
-        -------
-        tuple[ASE.Atoms, ASE.Atoms]
-            Tuple containing the two monomers.
-        """
-        if isinstance(atoms.info["selection_a"], str):
-            a_ids = [int(id) for id in atoms.info["selection_a"].split("-")]
-            a_ids[0] -= 1
-        else:
-            a_ids = [int(atoms.info["selection_a"]) - 1, int(atoms.info["selection_a"])]
-
-        if isinstance(atoms.info["selection_b"], str):
-            b_ids = [int(id) for id in atoms.info["selection_b"].split("-")]
-            b_ids[0] -= 1
-        else:
-            b_ids = [int(atoms.info["selection_b"]) - 1, int(atoms.info["selection_b"])]
-
-        atoms_a = atoms[a_ids[0] : a_ids[1]]
-        atoms_b = atoms[b_ids[0] : b_ids[1]]
-        assert len(atoms_a) + len(atoms_b) == len(atoms)
-
-        atoms_a.info["charge"] = int(atoms.info["charge_a"])
-        atoms_a.info["spin"] = 1
-
-        atoms_b.info["charge"] = int(atoms.info["charge_b"])
-        atoms_b.info["spin"] = 1
-        return (atoms_a, atoms_b)
-
-    def run(self):
-        """Run new benchmark."""
-        # Read in data and attach calculator
-        data_path = (
-            download_s3_data(
-                filename="NCIA_R739.zip",
-                key="inputs/non_covalent_interactions/NCIA_R739x5/NCIA_R739x5.zip",
-            )
-            / "NCIA_R739x5"
-        )
-        self.get_ref_energies(data_path)
-
-        calc = self.model.get_calculator()
-        # Add D3 calculator for this test
-        calc = self.model.add_d3_calculator(calc)
-
-        for label, ref_energy in tqdm(self.ref_energies.items()):
-            xyz_fname = f"{label}.xyz"
-            atoms = read(data_path / "geometries" / xyz_fname)
-            # exclude noble gases
-            if (
-                np.isin(atoms.numbers, [2, 10, 18, 36, 54, 86]).any()
-                and EXCLUDE_NOBLE_GASES
-            ):
+def get_ref_energies(data_path: Path) -> dict[str, float]:
+    """
+    Get reference energies.
+
+    Parameters
+    ----------
+    data_path
+        Path to data.
+
+    Returns
+    -------
+    dict[str, float]
+        Loaded reference energies.
+    """
+    ref_energies = {}
+
+    with open(data_path / "NCIA_R739x5_benchmark.txt") as lines:
+        for i, line in enumerate(lines):
+            if i < 3:
                 continue
-            atoms_a, atoms_b = self.get_monomers(atoms)
-            atoms.info["spin"] = 1
-            atoms.info["charge"] = int(atoms_a.info["charge"] + atoms_b.info["charge"])
-            atoms.calc = calc
-            atoms_a.calc = calc
-            atoms_b.calc = calc
-
-            atoms.info["model_int_energy"] = (
-                atoms.get_potential_energy()
-                - atoms_a.get_potential_energy()
-                - atoms_b.get_potential_energy()
-            )
-            atoms.info["ref_int_energy"] = ref_energy
-            atoms.calc = None
-
-            write_dir = OUT_PATH / self.model_name
-            write_dir.mkdir(parents=True, exist_ok=True)
-            write(write_dir / f"{label}.xyz", atoms)
-
-
-def build_project(repro: bool = False) -> None:
+            items = line.strip().split()
+            label = items[0]
+            ref_energy = float(items[1]) * KCAL_TO_EV
+            ref_energies[label] = ref_energy
+
+    return ref_energies
+
+
+def get_monomers(atoms: Atoms) -> tuple[Atoms, Atoms]:
     """
-    Build mlipx project.
+    Get ASE atoms objects of the monomers.
 
     Parameters
     ----------
-    repro
-        Whether to call dvc repro -f after building.
+    atoms
+        ASE atoms object of the structure.
+
+    Returns
+    -------
+    tuple[ASE.Atoms, ASE.Atoms]
+        Tuple containing the two monomers.
     """
-    project = mlipx.Project()
-    benchmark_node_dict = {}
-
-    for model_name, model in MODELS.items():
-        with project.group(model_name):
-            benchmark = NCIAR739x5Benchmark(
-                model=model,
-                model_name=model_name,
-            )
-            benchmark_node_dict[model_name] = benchmark
-
-    if repro:
-        with chdir(Path(__file__).parent):
-            project.repro(build=True, force=True)
+    if isinstance(atoms.info["selection_a"], str):
+        a_ids = [int(id) for id in atoms.info["selection_a"].split("-")]
+        a_ids[0] -= 1
+    else:
+        a_ids = [int(atoms.info["selection_a"]) - 1, int(atoms.info["selection_a"])]
+
+    if isinstance(atoms.info["selection_b"], str):
+        b_ids = [int(id) for id in atoms.info["selection_b"].split("-")]
+        b_ids[0] -= 1
     else:
-        project.build()
+        b_ids = [int(atoms.info["selection_b"]) - 1, int(atoms.info["selection_b"])]
+
+    atoms_a = atoms[a_ids[0] : a_ids[1]]
+    atoms_b = atoms[b_ids[0] : b_ids[1]]
+    assert len(atoms_a) + len(atoms_b) == len(atoms)
+
+    atoms_a.info["charge"] = int(atoms.info["charge_a"])
+    atoms_a.info["spin"] = 1
+
+    atoms_b.info["charge"] = int(atoms.info["charge_b"])
+    atoms_b.info["spin"] = 1
+    return (atoms_a, atoms_b)
+
+
+@pytest.mark.parametrize("mlip", MODELS.items())
+def test_ncia_r739x5(mlip: tuple[str, Any]) -> None:
+    """
+    Run NCIA R739x5 barriers benchmark.
 
+    Parameters
+    ----------
+    mlip
+        Name of model use and model to get calculator.
+    """
+    model_name, model = mlip
+    calc = model.get_calculator()
+
+    data_path = (
+        download_s3_data(
+            filename="NCIA_R739.zip",
+            key="inputs/non_covalent_interactions/NCIA_R739x5/NCIA_R739x5.zip",
+        )
+        / "NCIA_R739x5"
+    )
+    ref_energies = get_ref_energies(data_path)
+
+    calc = model.get_calculator()
+    # Add D3 calculator for this test
+    calc = model.add_d3_calculator(calc)
+
+    for label, ref_energy in tqdm(ref_energies.items()):
+        xyz_fname = f"{label}.xyz"
+        atoms = read(data_path / "geometries" / xyz_fname)
+        # exclude noble gases
+        if (
+            np.isin(atoms.numbers, [2, 10, 18, 36, 54, 86]).any()
+            and EXCLUDE_NOBLE_GASES
+        ):
+            continue
+        atoms_a, atoms_b = get_monomers(atoms)
+        atoms.info["spin"] = 1
+        atoms.info["charge"] = int(atoms_a.info["charge"] + atoms_b.info["charge"])
+        atoms.calc = calc
+        atoms_a.calc = calc
+        atoms_b.calc = calc
+
+        atoms.info["model_int_energy"] = (
+            atoms.get_potential_energy()
+            - atoms_a.get_potential_energy()
+            - atoms_b.get_potential_energy()
+        )
+        atoms.info["ref_int_energy"] = ref_energy
+        atoms.calc = None
 
-def test_ncia_r739x5():
-    """Run NCIA_R739x5 barriers benchmark via Pytest."""
-    build_project(repro=True)
+        write_dir = OUT_PATH / model_name
+        write_dir.mkdir(parents=True, exist_ok=True)
+        write(write_dir / f"{label}.xyz", atoms)

From 15eab0fd6ceef10ff4c170f93674aac865f1c5ee Mon Sep 17 00:00:00 2001
From: ElliottKasoar <45317199+ElliottKasoar@users.noreply.github.com>
Date: Fri, 30 Jan 2026 15:18:09 +0000
Subject: [PATCH 7/9] Rename benchmark

---
 .../analyse_NCIA_R739x5.py}                                    | 0
 .../{ncia_r739x5 => NCIA_R739x5}/metrics.yml                   | 0
 .../app_ncia_r739x5.py => NCIA_R739x5/app_NCIA_R739x5.py}      | 0
 .../calc_ncia_r739x5.py => NCIA_R739x5/calc_NCIA_R739x5.py}    | 0
 .../non_covalent_interactions/ncia_r739x5/.dvc/.gitignore      | 3 ---
 ml_peg/calcs/non_covalent_interactions/ncia_r739x5/.dvc/config | 0
 ml_peg/calcs/non_covalent_interactions/ncia_r739x5/.dvcignore  | 3 ---
 7 files changed, 6 deletions(-)
 rename ml_peg/analysis/non_covalent_interactions/{ncia_r739x5/analyse_ncia_r739x5.py => NCIA_R739x5/analyse_NCIA_R739x5.py} (100%)
 rename ml_peg/analysis/non_covalent_interactions/{ncia_r739x5 => NCIA_R739x5}/metrics.yml (100%)
 rename ml_peg/app/non_covalent_interactions/{ncia_r739x5/app_ncia_r739x5.py => NCIA_R739x5/app_NCIA_R739x5.py} (100%)
 rename ml_peg/calcs/non_covalent_interactions/{ncia_r739x5/calc_ncia_r739x5.py => NCIA_R739x5/calc_NCIA_R739x5.py} (100%)
 delete mode 100644 ml_peg/calcs/non_covalent_interactions/ncia_r739x5/.dvc/.gitignore
 delete mode 100644 ml_peg/calcs/non_covalent_interactions/ncia_r739x5/.dvc/config
 delete mode 100644 ml_peg/calcs/non_covalent_interactions/ncia_r739x5/.dvcignore

diff --git a/ml_peg/analysis/non_covalent_interactions/ncia_r739x5/analyse_ncia_r739x5.py b/ml_peg/analysis/non_covalent_interactions/NCIA_R739x5/analyse_NCIA_R739x5.py
similarity index 100%
rename from ml_peg/analysis/non_covalent_interactions/ncia_r739x5/analyse_ncia_r739x5.py
rename to ml_peg/analysis/non_covalent_interactions/NCIA_R739x5/analyse_NCIA_R739x5.py
diff --git a/ml_peg/analysis/non_covalent_interactions/ncia_r739x5/metrics.yml b/ml_peg/analysis/non_covalent_interactions/NCIA_R739x5/metrics.yml
similarity index 100%
rename from ml_peg/analysis/non_covalent_interactions/ncia_r739x5/metrics.yml
rename to ml_peg/analysis/non_covalent_interactions/NCIA_R739x5/metrics.yml
diff --git a/ml_peg/app/non_covalent_interactions/ncia_r739x5/app_ncia_r739x5.py b/ml_peg/app/non_covalent_interactions/NCIA_R739x5/app_NCIA_R739x5.py
similarity index 100%
rename from ml_peg/app/non_covalent_interactions/ncia_r739x5/app_ncia_r739x5.py
rename to ml_peg/app/non_covalent_interactions/NCIA_R739x5/app_NCIA_R739x5.py
diff --git a/ml_peg/calcs/non_covalent_interactions/ncia_r739x5/calc_ncia_r739x5.py b/ml_peg/calcs/non_covalent_interactions/NCIA_R739x5/calc_NCIA_R739x5.py
similarity index 100%
rename from ml_peg/calcs/non_covalent_interactions/ncia_r739x5/calc_ncia_r739x5.py
rename to ml_peg/calcs/non_covalent_interactions/NCIA_R739x5/calc_NCIA_R739x5.py
diff --git a/ml_peg/calcs/non_covalent_interactions/ncia_r739x5/.dvc/.gitignore b/ml_peg/calcs/non_covalent_interactions/ncia_r739x5/.dvc/.gitignore
deleted file mode 100644
index 528f30c71..000000000
--- a/ml_peg/calcs/non_covalent_interactions/ncia_r739x5/.dvc/.gitignore
+++ /dev/null
@@ -1,3 +0,0 @@
-/config.local
-/tmp
-/cache
diff --git a/ml_peg/calcs/non_covalent_interactions/ncia_r739x5/.dvc/config b/ml_peg/calcs/non_covalent_interactions/ncia_r739x5/.dvc/config
deleted file mode 100644
index e69de29bb..000000000
diff --git a/ml_peg/calcs/non_covalent_interactions/ncia_r739x5/.dvcignore b/ml_peg/calcs/non_covalent_interactions/ncia_r739x5/.dvcignore
deleted file mode 100644
index 519730552..000000000
--- a/ml_peg/calcs/non_covalent_interactions/ncia_r739x5/.dvcignore
+++ /dev/null
@@ -1,3 +0,0 @@
-# Add patterns of files dvc should ignore, which could improve
-# the performance. Learn more at
-# https://dvc.org/doc/user-guide/dvcignore

From d3ef8c99e73c8db89fd2071cd4763e724ce05754 Mon Sep 17 00:00:00 2001
From: ElliottKasoar <45317199+ElliottKasoar@users.noreply.github.com>
Date: Fri, 30 Jan 2026 15:23:06 +0000
Subject: [PATCH 8/9] Tidy and use kcal/mol

---
 .../NCIA_R739x5/analyse_NCIA_R739x5.py        | 19 +++++++++----------
 .../NCIA_R739x5/metrics.yml                   |  4 ++--
 .../NCIA_R739x5/app_NCIA_R739x5.py            |  6 +++---
 .../NCIA_R739x5/calc_NCIA_R739x5.py           |  1 -
 4 files changed, 14 insertions(+), 16 deletions(-)

diff --git a/ml_peg/analysis/non_covalent_interactions/NCIA_R739x5/analyse_NCIA_R739x5.py b/ml_peg/analysis/non_covalent_interactions/NCIA_R739x5/analyse_NCIA_R739x5.py
index 7ebc5e467..d78f8bbe1 100644
--- a/ml_peg/analysis/non_covalent_interactions/NCIA_R739x5/analyse_NCIA_R739x5.py
+++ b/ml_peg/analysis/non_covalent_interactions/NCIA_R739x5/analyse_NCIA_R739x5.py
@@ -1,4 +1,4 @@
-"""Analyse ncia_r739x5 benchmark."""
+"""Analyse NCIA R739x5 benchmark."""
 
 from __future__ import annotations
 
@@ -18,10 +18,9 @@
 MODELS = load_models(current_models)
 D3_MODEL_NAMES = build_d3_name_map(MODELS)
 
-KCAL_TO_EV = units.kcal / units.mol
-EV_TO_KCAL = 1 / KCAL_TO_EV
-CALC_PATH = CALCS_ROOT / "non_covalent_interactions" / "ncia_r739x5" / "outputs"
-OUT_PATH = APP_ROOT / "data" / "non_covalent_interactions" / "ncia_r739x5"
+EV_TO_KCAL = units.mol / units.kcal
+CALC_PATH = CALCS_ROOT / "non_covalent_interactions" / "NCIA_R739x5" / "outputs"
+OUT_PATH = APP_ROOT / "data" / "non_covalent_interactions" / "NCIA_R739x5"
 
 METRICS_CONFIG_PATH = Path(__file__).with_name("metrics.yml")
 DEFAULT_THRESHOLDS, DEFAULT_TOOLTIPS, DEFAULT_WEIGHTS = load_metrics_config(
@@ -48,8 +47,8 @@ def labels() -> list:
 @plot_parity(
     filename=OUT_PATH / "figure_ncia_r739x5.json",
     title="Interaction energies",
-    x_label="Predicted energy / eV",
-    y_label="Reference energy / eV",
+    x_label="Predicted energy / kcal/mol",
+    y_label="Reference energy / kcal/mol",
     hoverdata={
         "Labels": labels(),
     },
@@ -71,9 +70,9 @@ def interaction_energies() -> dict[str, list]:
         for label in labels():
             atoms = read(CALC_PATH / model_name / f"{label}.xyz")
             if not ref_stored:
-                results["ref"].append(atoms.info["ref_int_energy"])
+                results["ref"].append(atoms.info["ref_int_energy"] * EV_TO_KCAL)
 
-            results[model_name].append(atoms.info["model_int_energy"])
+            results[model_name].append(atoms.info["model_int_energy"] * EV_TO_KCAL)
 
             # Write structures for app
             structs_dir = OUT_PATH / model_name
@@ -135,7 +134,7 @@ def metrics(get_mae: dict[str, float]) -> dict[str, dict]:
 
 def test_ncia_r739x5(metrics: dict[str, dict]) -> None:
     """
-    Run ncia_r739x5 test.
+    Run NCIA R739x5 test.
 
     Parameters
     ----------
diff --git a/ml_peg/analysis/non_covalent_interactions/NCIA_R739x5/metrics.yml b/ml_peg/analysis/non_covalent_interactions/NCIA_R739x5/metrics.yml
index 8118e8238..45388b831 100644
--- a/ml_peg/analysis/non_covalent_interactions/NCIA_R739x5/metrics.yml
+++ b/ml_peg/analysis/non_covalent_interactions/NCIA_R739x5/metrics.yml
@@ -1,7 +1,7 @@
 metrics:
   MAE:
     good: 0.0
-    bad: 0.5
-    unit: eV
+    bad: 10
+    unit: kcal/mol
     tooltip: Mean Absolute Error for all systems
     level_of_theory: CCSD(T)
diff --git a/ml_peg/app/non_covalent_interactions/NCIA_R739x5/app_NCIA_R739x5.py b/ml_peg/app/non_covalent_interactions/NCIA_R739x5/app_NCIA_R739x5.py
index faee0c4dd..f79e564c9 100644
--- a/ml_peg/app/non_covalent_interactions/NCIA_R739x5/app_NCIA_R739x5.py
+++ b/ml_peg/app/non_covalent_interactions/NCIA_R739x5/app_NCIA_R739x5.py
@@ -16,12 +16,12 @@
 from ml_peg.models.models import current_models
 
 MODELS = get_model_names(current_models)
-BENCHMARK_NAME = "NCIA_R739x5"
+BENCHMARK_NAME = "NCIA R739x5"
 DOCS_URL = (
     "https://ddmms.github.io/ml-peg/user_guide/benchmarks/"
     "non_covalent_interactions.html#ncia-r739x5"
 )
-DATA_PATH = APP_ROOT / "data" / "non_covalent_interactions" / "ncia_r739x5"
+DATA_PATH = APP_ROOT / "data" / "non_covalent_interactions" / "NCIA_R739x5"
 
 
 class NCIAR739x5App(BaseApp):
@@ -38,7 +38,7 @@ def register_callbacks(self) -> None:
         if model_dir.exists():
             labels = sorted([f.stem for f in model_dir.glob("*.xyz")])
             structs = [
-                f"assets/non_covalent_interactions/ncia_r739x5/{MODELS[0]}/{label}.xyz"
+                f"assets/non_covalent_interactions/NCIA_R739x5/{MODELS[0]}/{label}.xyz"
                 for label in labels
             ]
         else:
diff --git a/ml_peg/calcs/non_covalent_interactions/NCIA_R739x5/calc_NCIA_R739x5.py b/ml_peg/calcs/non_covalent_interactions/NCIA_R739x5/calc_NCIA_R739x5.py
index 18d02e675..98aa7c363 100644
--- a/ml_peg/calcs/non_covalent_interactions/NCIA_R739x5/calc_NCIA_R739x5.py
+++ b/ml_peg/calcs/non_covalent_interactions/NCIA_R739x5/calc_NCIA_R739x5.py
@@ -22,7 +22,6 @@
 MODELS = load_models(current_models)
 
 KCAL_TO_EV = units.kcal / units.mol
-EV_TO_KCAL = 1 / KCAL_TO_EV
 
 OUT_PATH = Path(__file__).parent / "outputs"
 EXCLUDE_NOBLE_GASES = True

From f226dc6c1c5fbb23700409b9402bb0fd6bcd7246 Mon Sep 17 00:00:00 2001
From: joehart2001 <jh2536@cam.ac.uk>
Date: Wed, 4 Feb 2026 16:01:47 +0000
Subject: [PATCH 9/9] add density scatter

---
 .../NCIA_R739x5/analyse_NCIA_R739x5.py        | 55 +++++++++++++++----
 .../NCIA_R739x5/app_NCIA_R739x5.py            | 42 +++++---------
 2 files changed, 57 insertions(+), 40 deletions(-)

diff --git a/ml_peg/analysis/non_covalent_interactions/NCIA_R739x5/analyse_NCIA_R739x5.py b/ml_peg/analysis/non_covalent_interactions/NCIA_R739x5/analyse_NCIA_R739x5.py
index d78f8bbe1..333e01f2e 100644
--- a/ml_peg/analysis/non_covalent_interactions/NCIA_R739x5/analyse_NCIA_R739x5.py
+++ b/ml_peg/analysis/non_covalent_interactions/NCIA_R739x5/analyse_NCIA_R739x5.py
@@ -8,7 +8,10 @@
 from ase.io import read, write
 import pytest
 
-from ml_peg.analysis.utils.decorators import build_table, plot_parity
+from ml_peg.analysis.utils.decorators import (
+    build_table,
+    plot_density_scatter,
+)
 from ml_peg.analysis.utils.utils import build_d3_name_map, load_metrics_config, mae
 from ml_peg.app import APP_ROOT
 from ml_peg.calcs import CALCS_ROOT
@@ -44,15 +47,6 @@ def labels() -> list:
 
 
 @pytest.fixture
-@plot_parity(
-    filename=OUT_PATH / "figure_ncia_r739x5.json",
-    title="Interaction energies",
-    x_label="Predicted energy / kcal/mol",
-    y_label="Reference energy / kcal/mol",
-    hoverdata={
-        "Labels": labels(),
-    },
-)
 def interaction_energies() -> dict[str, list]:
     """
     Get interaction energies for all systems.
@@ -83,6 +77,40 @@ def interaction_energies() -> dict[str, list]:
     return results
 
 
+@pytest.fixture
+@plot_density_scatter(
+    filename=OUT_PATH / "figure_ncia_r739x5_density.json",
+    title="Interaction energy density plot",
+    x_label="Reference energy / kcal/mol",
+    y_label="Predicted energy / kcal/mol",
+    annotation_metadata={"system_count": "Systems"},
+)
+def interaction_density(interaction_energies: dict[str, list]) -> dict[str, dict]:
+    """
+    Build density scatter inputs for interaction energies.
+
+    Parameters
+    ----------
+    interaction_energies
+        Reference and predicted interaction energies per model.
+
+    Returns
+    -------
+    dict[str, dict]
+        Mapping of model names to density-plot payloads.
+    """
+    ref_vals = interaction_energies["ref"]
+    density_inputs: dict[str, dict] = {}
+    for model_name in MODELS:
+        preds = interaction_energies.get(model_name, [])
+        density_inputs[model_name] = {
+            "ref": ref_vals,
+            "pred": preds,
+            "meta": {"system_count": len([val for val in preds if val is not None])},
+        }
+    return density_inputs
+
+
 @pytest.fixture
 def get_mae(interaction_energies) -> dict[str, float]:
     """
@@ -132,7 +160,10 @@ def metrics(get_mae: dict[str, float]) -> dict[str, dict]:
     }
 
 
-def test_ncia_r739x5(metrics: dict[str, dict]) -> None:
+def test_ncia_r739x5(
+    metrics: dict[str, dict],
+    interaction_density: dict[str, dict],
+) -> None:
     """
     Run NCIA R739x5 test.
 
@@ -140,5 +171,7 @@ def test_ncia_r739x5(metrics: dict[str, dict]) -> None:
     ----------
     metrics
         All new benchmark metric names and dictionary of values for each model.
+    interaction_density
+        Density-scatter inputs for all models (drives saved plots).
     """
     return
diff --git a/ml_peg/app/non_covalent_interactions/NCIA_R739x5/app_NCIA_R739x5.py b/ml_peg/app/non_covalent_interactions/NCIA_R739x5/app_NCIA_R739x5.py
index f79e564c9..f9ea4f6c6 100644
--- a/ml_peg/app/non_covalent_interactions/NCIA_R739x5/app_NCIA_R739x5.py
+++ b/ml_peg/app/non_covalent_interactions/NCIA_R739x5/app_NCIA_R739x5.py
@@ -7,11 +7,8 @@
 
 from ml_peg.app import APP_ROOT
 from ml_peg.app.base_app import BaseApp
-from ml_peg.app.utils.build_callbacks import (
-    plot_from_table_column,
-    struct_from_scatter,
-)
-from ml_peg.app.utils.load import read_plot
+from ml_peg.app.utils.build_callbacks import plot_from_table_cell
+from ml_peg.app.utils.load import read_density_plot_for_model
 from ml_peg.models.get_models import get_model_names
 from ml_peg.models.models import current_models
 
@@ -29,32 +26,20 @@ class NCIAR739x5App(BaseApp):
 
     def register_callbacks(self) -> None:
         """Register callbacks to app."""
-        scatter = read_plot(
-            DATA_PATH / "figure_ncia_r739x5.json",
-            id=f"{BENCHMARK_NAME}-figure",
-        )
-
-        model_dir = DATA_PATH / MODELS[0]
-        if model_dir.exists():
-            labels = sorted([f.stem for f in model_dir.glob("*.xyz")])
-            structs = [
-                f"assets/non_covalent_interactions/NCIA_R739x5/{MODELS[0]}/{label}.xyz"
-                for label in labels
-            ]
-        else:
-            structs = []
+        density_plots: dict[str, dict] = {}
+        for model in MODELS:
+            density_graph = read_density_plot_for_model(
+                filename=DATA_PATH / "figure_ncia_r739x5_density.json",
+                model=model,
+                id=f"{BENCHMARK_NAME}-{model}-density",
+            )
+            if density_graph is not None:
+                density_plots[model] = {"MAE": density_graph}
 
-        plot_from_table_column(
+        plot_from_table_cell(
             table_id=self.table_id,
             plot_id=f"{BENCHMARK_NAME}-figure-placeholder",
-            column_to_plot={"MAE": scatter},
-        )
-
-        struct_from_scatter(
-            scatter_id=f"{BENCHMARK_NAME}-figure",
-            struct_id=f"{BENCHMARK_NAME}-struct-placeholder",
-            structs=structs,
-            mode="struct",
+            cell_to_plot=density_plots,
         )
 
 
@@ -78,7 +63,6 @@ def get_app() -> NCIAR739x5App:
         table_path=DATA_PATH / "ncia_r739x5_metrics_table.json",
         extra_components=[
             Div(id=f"{BENCHMARK_NAME}-figure-placeholder"),
-            Div(id=f"{BENCHMARK_NAME}-struct-placeholder"),
         ],
     )