Advanced Plotting¶

Once you’re familiar with RISK’s core plotting methods, you can start layering annotations, subnetworks, and styles to highlight specific biological themes. This section walks through advanced features using a real-world example.

Use Case: Dual Overlay of Coexpression Subgroups¶

In this example, we highlight two functionally distinct sets of genes:

Highly coexpressed nodes
Weakly coexpressed nodes

We annotate each with a distinct color and size overlay, draw the global network structure, and apply clear labeling for biological interpretation.

Step 1: Define Your Subgroups¶

Use lists of node names to define subgroups of interest.

high_coexpression_nodes = ["RPL1A", "RPL2B", "RPL3", "RPL5"]
low_coexpression_nodes = ["YGR103W", "YDL064W", "YML082W", "YOR204W"]

Step 2: Initialize the Plotter¶

Before plotting, ensure that the graph object has been instantiated. This should be the same NetworkGraph object created during the setup in the Load Graph section. The graph instance provides the structural and annotation context necessary for visual overlays such as coexpression subnetworks or GO BP contours.

We create a new NetworkPlotter instance and set background preferences.

plotter = risk.load_plotter(
    graph=graph,
    figsize=(16, 16),
    background_color="black",
    background_alpha=1.0,
    pad=0.3,
)

Step 3: Draw the Full Network and Frame¶

plotter.plot_circle_perimeter(
    radius=1.0,
    color="white",
    linewidth=2.0,
    linestyle="solid",
    alpha=1.0,
)

plotter.plot_network(
    node_colors=plotter.get_annotated_node_colors(),
    node_sizes=plotter.get_annotated_node_sizes(),
)

Step 4: Highlight Subnetworks¶

Overlay the high and low coexpression groups in distinct styles.

plotter.plot_subnetwork(
    nodes=high_coexpression_nodes,
    color="lime",
    size=225,
    alpha=1.0,
)

plotter.plot_subnetwork(
    nodes=low_coexpression_nodes,
    color="magenta",
    size=225,
    alpha=1.0,
)

Step 5: Add KDE Domain Contours¶

Highlight overrepresented domains from graph using KDE-based clustering.

plotter.plot_contours(
    levels=5,
    grid_size=250,
    color=plotter.get_annotated_contour_colors(),
    alpha=1.0,
    fill_alpha=0.25,
)

Step 6: Annotate Regions with Sublabels¶

Draw manual labels and arrows pointing to desired subnetwork regions.

plotter.plot_sublabel(
    nodes=high_coexpression_nodes,
    label="High Coexpression (Ribosomal)",
    radial_position=45,
    scale=1.8,
    offset=0.12,
    fontcolor="lime",
    fontsize=16,
)

plotter.plot_sublabel(
    nodes=low_coexpression_nodes,
    label="Low Coexpression (Metabolic)",
    radial_position=190,
    scale=1.8,
    offset=0.12,
    fontcolor="magenta",
    fontsize=16,
)

Use Case: Highlighting Party Hubs and Date Hubs¶

In this advanced example, we visualize party hubs and date hubs based on coexpression scores from the SPELL database. We use precomputed color mappings for each subgroup and annotate regions of the network linked to ribosomal and metabolic functions.

Step 1: Load and Plot Colored Subnetworks¶

import json

def load_json_to_dict(filepath):
    with open(filepath, "r") as file:
        return json.load(file)

high_rgba = load_json_to_dict("./data/json/coexpression/high_coexpression_michaelis_2023.json")
low_rgba = load_json_to_dict("./data/json/coexpression/low_coexpression_michaelis_2023.json")

high_nodes, high_colors = zip(*high_rgba.items())
low_nodes, low_colors = zip(*low_rgba.items())

plotter = risk.load_plotter(graph=graph, figsize=(15, 15), background_color="black")

plotter.plot_subnetwork(nodes=high_nodes, node_size=1500, node_color=high_colors, edge_width=0)
plotter.plot_subnetwork(nodes=low_nodes, node_size=1500, node_color=low_colors, edge_width=0)
plotter.plot_contours(color="white", alpha=0.75, linewidth=2.5)

Step 2: Annotate Ribosomal and Metabolic Domains¶

Use domain IDs (discovered via plot_labels(overlay_ids=True)) to retrieve node labels for key pathways.

def get_labels(domain_ids):
    if not isinstance(domain_ids, (list, tuple)):
        domain_ids = [domain_ids]
    return sum([graph.domain_id_to_node_labels_map[d] for d in domain_ids], [])

# Note: These domain IDs are based on a specific random seed. They may differ if layout or clustering is reinitialized.
ribosome_nodes = get_labels([69, 72, 28, 79])

# Domain IDs for metabolic processes; adjust as needed if using a different seed or graph instance.
metabolism_nodes = get_labels([16, 18, 41, 99])

plotter.plot_sublabel(nodes=ribosome_nodes, label="Ribosomal\nProcesses", radial_position=240)
plotter.plot_sublabel(nodes=metabolism_nodes, label="Metabolic\nProcesses", radial_position=60)
plotter.show()

This overlay highlights the spatial and functional separation of coexpression-based hub classes in the yeast PPI network.

Final Output¶

Display or save the composite figure.

plotter.savefig("advanced_plot.png", dpi=300)
plotter.show()

Party & Date Hubs

Recap: What You Learned¶

How to highlight custom biological subgroups
How to blend annotation significance with custom overlays
How to label specific domains using radial positioning
How to layer multiple visual features for storytelling

Next Step¶

Explore example notebooks or 8. Parameters for a complete list of available options.