cytocommunity2 0.1.6__tar.gz → 0.1.7__tar.gz

{cytocommunity2-0.1.6/src/cytocommunity2.egg-info → cytocommunity2-0.1.7}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: cytocommunity2
-Version: 0.1.6
+Version: 0.1.7
 Summary: Weakly supervised cellular-neighborhood learning for spatial omics
 Requires-Python: >=3.10
 Description-Content-Type: text/markdown

{cytocommunity2-0.1.6 → cytocommunity2-0.1.7}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "cytocommunity2"
-version = "0.1.6"
+version = "0.1.7"
 description = "Weakly supervised cellular-neighborhood learning for spatial omics"
 readme = "README.md"
 requires-python = ">=3.10"

{cytocommunity2-0.1.6 → cytocommunity2-0.1.7}/src/cytocommunity2/__init__.py

@@ -10,7 +10,7 @@ from .learning import (
 )
 from .visualization import visualize_results
 
-__version__ = "0.1.6"
+__version__ = "0.1.7"
 
 __all__ = [
     "CytoCommunityConfig",

{cytocommunity2-0.1.6 → cytocommunity2-0.1.7}/src/cytocommunity2/downstream/plotting/dominant.py

@@ -53,10 +53,7 @@ def plot_dominant_celltype_results(result, config):
         label.set_horizontalalignment("right")
     heat_axis.tick_params(axis="x", length=0, bottom=True, top=False, labelbottom=True)
     heat_axis.tick_params(axis="y", length=0, pad=6)
-    heat_axis.set_xticks(np.arange(-0.5, columns, 1), minor=True)
-    heat_axis.set_yticks(np.arange(-0.5, rows, 1), minor=True)
-    heat_axis.grid(which="minor", color="white", linewidth=0.35)
-    heat_axis.tick_params(which="minor", bottom=False, left=False)
+    heat_axis.grid(False, which="both")
     for spine in heat_axis.spines.values():
         spine.set_visible(False)
     heatmap_output = output_dir / "heatmap_main.pdf"

{cytocommunity2-0.1.6 → cytocommunity2-0.1.7}/src/cytocommunity2/downstream/plotting/recurrence.py

@@ -8,6 +8,8 @@ from ...paths import build_paths
 
 NULL_LABEL = "Random-CN null distribution"
 NULL_COLOR = "#0071B2"
+CN_PALETTE = "d3_category20"
+DENSITY_NULL_COLOR = "#B8B8B8"
 
 
 def _significance_label(pvalue):
@@ -28,11 +30,11 @@ def _kde_curve(values, x_grid):
     if len(values) == 0:
         return np.zeros_like(x_grid)
     if np.allclose(values, values[0]):
-        density = np.zeros_like(x_grid)
-        density[np.argmin(np.abs(x_grid - values[0]))] = 1.0
-        density = np.convolve(density, np.array([0.2, 0.6, 0.2]), mode="same")
-        area = np.trapz(density, x_grid)
-        return density / area if area > 0 else density
+        # gaussian_kde cannot estimate a singular distribution. Draw a narrow,
+        # finite kernel so a constant-valued CN still remains visible.
+        bandwidth = 0.04
+        scaled = (x_grid - values[0]) / bandwidth
+        return np.exp(-0.5 * scaled**2) / (bandwidth * np.sqrt(2 * np.pi))
     try:
         return gaussian_kde(values)(x_grid)
     except Exception:
@@ -40,8 +42,21 @@ def _kde_curve(values, x_grid):
         return np.interp(x_grid, (edges[:-1] + edges[1:]) / 2, hist, left=0, right=0)
 
 
+def _cn_color_map(cn_levels):
+    """Return the same CN-to-color mapping used by the spatial CN plots."""
+    import sci_palettes
+
+    colors = list(sci_palettes.PALETTES[CN_PALETTE].values())
+    if len(cn_levels) > len(colors):
+        raise ValueError(
+            f"Palette {CN_PALETTE!r} has {len(colors)} colors, but "
+            f"{len(cn_levels)} CNs were provided."
+        )
+    return dict(zip(cn_levels, colors))
+
+
 def plot_recurrence_results(config):
-    """Save the three original recurrence violin/KDE figures."""
+    """Save recurrence violin/KDE figures, including per-CN densities."""
     import matplotlib as mpl
     import matplotlib.pyplot as plt
     import seaborn as sns
@@ -196,8 +211,63 @@ def plot_recurrence_results(config):
     kde_output = output_dir / "Overall_Real_vs_randomCN_null_KDE.pdf"
     figure.savefig(kde_output, dpi=1200)
     plt.close(figure)
+
+    cn_colors = _cn_color_map(cn_levels)
+    null_density = _kde_curve(background_values, x_grid)
+    figure, axis = plt.subplots(figsize=(8.4, 5.8))
+    axis.fill_between(
+        x_grid, 0, null_density, color=DENSITY_NULL_COLOR, alpha=0.18, zorder=0
+    )
+    axis.plot(
+        x_grid,
+        null_density,
+        color=DENSITY_NULL_COLOR,
+        linewidth=2.2,
+        linestyle="--",
+        label="Null distribution",
+        zorder=1,
+    )
+    for cn in cn_levels:
+        values = real.loc[real["CN"] == cn, "OverlapCoefficient"].to_numpy()
+        density = _kde_curve(values, x_grid)
+        color = cn_colors[cn]
+        axis.fill_between(x_grid, 0, density, color=color, alpha=0.09, zorder=2)
+        axis.plot(x_grid, density, color=color, linewidth=2.0, label=cn, zorder=3)
+
+    axis.set_xlabel("Overlap of dominant celltypes across samples")
+    axis.set_ylabel("Density")
+    axis.set_xlim(0, 1)
+    axis.set_ylim(bottom=0)
+    axis.spines["top"].set_visible(False)
+    axis.spines["right"].set_visible(False)
+    handles, labels = axis.get_legend_handles_labels()
+    legend_columns = min(6, len(labels))
+    legend_order = [
+        index
+        for column in range(legend_columns)
+        for index in (column, legend_columns + column)
+        if index < len(labels)
+    ]
+    axis.legend(
+        [handles[index] for index in legend_order],
+        [labels[index] for index in legend_order],
+        frameon=False,
+        ncol=legend_columns,
+        loc="upper center",
+        bbox_to_anchor=(0.5, -0.17),
+        columnspacing=1.25,
+        handlelength=2.0,
+    )
+    figure.subplots_adjust(left=0.12, right=0.98, bottom=0.25, top=0.96)
+    by_cn_density_output = output_dir / "OC_density_by_CN_with_randomCN_null.pdf"
+    figure.savefig(by_cn_density_output, dpi=1200, bbox_inches="tight")
+    figure.savefig(
+        by_cn_density_output.with_suffix(".png"), dpi=300, bbox_inches="tight"
+    )
+    plt.close(figure)
     return {
         "by_cn": detailed_output,
         "pooled": pooled_output,
         "kde": kde_output,
+        "by_cn_density": by_cn_density_output,
     }

{cytocommunity2-0.1.6 → cytocommunity2-0.1.7/src/cytocommunity2.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: cytocommunity2
-Version: 0.1.6
+Version: 0.1.7
 Summary: Weakly supervised cellular-neighborhood learning for spatial omics
 Requires-Python: >=3.10
 Description-Content-Type: text/markdown