Specialized Visualization

This document provides comprehensive documentation for specialized visualization types in METAINFORMANT, including Venn diagrams, Sankey diagrams, chord diagrams, alluvial plots, and other advanced chart types used in bioinformatics.

Overview

Specialized visualization includes advanced plotting types commonly used in systems biology, set analysis, flow visualization, and complex data relationships. These tools complement the core plotting modules with publication-quality specialized charts.

Module Functions

Set Analysis

Venn Diagrams

from metainformant.visualization import specialized as spec_viz

# Create Venn diagram for set intersections
sets = {
    'Dataset_A': {'gene1', 'gene2', 'gene3', 'gene4'},
    'Dataset_B': {'gene2', 'gene3', 'gene5', 'gene6'},
    'Dataset_C': {'gene1', 'gene3', 'gene5', 'gene7'}
}

ax = spec_viz.plot_venn_diagram(sets, figsize=(8, 8))

UpSet Plots

# Create UpSet plot for complex set intersections
data = {
    'Condition_1': {'A', 'B', 'C', 'D'},
    'Condition_2': {'B', 'C', 'E', 'F'},
    'Condition_3': {'A', 'C', 'E', 'G'},
    'Condition_4': {'B', 'D', 'F', 'H'}
}

fig = spec_viz.plot_upset_plot(data, figsize=(10, 6))

Flow Visualization

Sankey Diagrams

# Create Sankey diagram for flow data
flows = [
    ('Source_A', 'Target_X', 10),
    ('Source_A', 'Target_Y', 5),
    ('Source_B', 'Target_X', 8),
    ('Source_B', 'Target_Z', 12),
    ('Source_C', 'Target_Y', 6)
]

# Static matplotlib version
ax = spec_viz.plot_sankey_diagram(flows, figsize=(10, 8))

# Interactive Plotly version
fig = spec_viz.plot_sankey_diagram(flows, figsize=(10, 8), interactive=True)

Alluvial Diagrams

# Create alluvial diagram for transitions
import pandas as pd

data = pd.DataFrame({
    'Stage_1': ['A', 'A', 'B', 'B', 'C', 'C'],
    'Stage_2': ['X', 'Y', 'X', 'Y', 'X', 'Z'],
    'Stage_3': ['P', 'P', 'Q', 'Q', 'R', 'R'],
    'Count': [5, 3, 4, 6, 2, 8]
})

ax = spec_viz.plot_alluvial_diagram(data, ['Stage_1', 'Stage_2', 'Stage_3'])

Network Analysis

Chord Diagrams

# Create chord diagram for relationships
matrix = np.array([
    [0, 5, 3, 2],
    [5, 0, 4, 1],
    [3, 4, 0, 6],
    [2, 1, 6, 0]
])
labels = ['Group_A', 'Group_B', 'Group_C', 'Group_D']

ax = spec_viz.plot_chord_diagram(matrix, labels, figsize=(8, 8))

Circular Network Layout

# Plot network in circular layout
import networkx as nx

G = nx.erdos_renyi_graph(20, 0.1)
ax = spec_viz.plot_network_circular_layout(G, figsize=(10, 10))

Multi-dimensional Data

Radar Charts

# Create radar chart for multi-dimensional data
import pandas as pd

data = pd.DataFrame({
    'Sample_1': [0.8, 0.6, 0.9, 0.4, 0.7],
    'Sample_2': [0.6, 0.8, 0.5, 0.6, 0.8],
    'Sample_3': [0.7, 0.7, 0.7, 0.7, 0.7]
}, index=['Trait_A', 'Trait_B', 'Trait_C', 'Trait_D', 'Trait_E'])

ax = spec_viz.plot_radar_chart(data, figsize=(8, 8))

Circular Bar Plots

# Create circular bar plot
values = np.random.rand(12) * 100
labels = [f'Category_{i+1}' for i in range(12)]

ax = spec_viz.plot_circular_barplot(values, labels, figsize=(8, 8))

Integration Examples

With Multi-Omics Data

from metainformant.multiomics import integration, visualization as multi_viz
from metainformant.visualization import specialized as spec_viz

# Integrate multi-omics data
omics_data = {
    'rna': pd.DataFrame(np.random.rand(100, 50)),
    'protein': pd.DataFrame(np.random.rand(100, 30)),
    'metabolomics': pd.DataFrame(np.random.rand(100, 20))
}

integrated = integration.integrate_omics_data(**omics_data)

# Create specialized visualizations
# UpSet plot for feature overlaps
feature_sets = {
    'RNA': set(range(50)),
    'Protein': set(range(50, 80)),
    'Metabolites': set(range(80, 100))
}
fig = spec_viz.plot_upset_plot(feature_sets)

With GWAS Results

from metainformant.gwas import visualization as gwas_viz
from metainformant.visualization import specialized as spec_viz

# GWAS analysis results
# Create Sankey diagram for workflow steps
workflow_flows = [
    ('Raw_Variants', 'QC_Filtered', 950000),
    ('Raw_Variants', 'Removed_LowQual', 50000),
    ('QC_Filtered', 'Associated_SNPs', 5000),
    ('QC_Filtered', 'Non_Associated', 945000),
    ('Associated_SNPs', 'Significant_Hits', 500),
    ('Associated_SNPs', 'Suggestive_Hits', 4500)
]

fig = spec_viz.plot_sankey_diagram(workflow_flows)

With Ontology Analysis

from metainformant.ontology.core import go
from metainformant.ontology.visualization import visualization as ont_viz
from metainformant.visualization import specialized as spec_viz

# GO enrichment results
# Create chord diagram for GO term relationships
go_terms = ['GO:0008150', 'GO:0003674', 'GO:0005575', 'GO:0009987']
relationship_matrix = np.random.rand(4, 4)
relationship_matrix = (relationship_matrix + relationship_matrix.T) / 2
np.fill_diagonal(relationship_matrix, 0)

ax = spec_viz.plot_chord_diagram(relationship_matrix, go_terms)

Output Options

All specialized visualization functions support:

# Save static plots
ax = spec_viz.plot_venn_diagram(sets, output_path="venn_diagram.png")

# Save interactive plots
fig = spec_viz.plot_sankey_diagram(flows, output_path="sankey.html")

# UpSet plots
fig = spec_viz.plot_upset_plot(data, output_path="upset.png")

Dependencies

Required: matplotlib, numpy
Venn Diagrams: matplotlib-venn (optional, fallback provided)
UpSet Plots: upsetplot (optional)
Alluvial Diagrams: alluvial (optional)
Interactive Plots: plotly (optional)
Network Analysis: networkx (optional)

Performance Considerations

Large Datasets: Venn diagrams work best with <5 sets
Complex Networks: Circular layouts scale to ~100 nodes
Interactive Plots: Plotly Sankey diagrams handle thousands of flows
Memory Usage: Chord diagrams require O(n²) memory for dense matrices

Color Schemes

Recommended color schemes for specialized plots:

# Venn diagrams
venn_colors = ['#FF6B6B', '#4ECDC4', '#45B7D1', '#96CEB4', '#FFEAA7']

# Sankey diagrams
sankey_colors = {
    'source': '#FF6B6B',
    'target': '#4ECDC4',
    'intermediate': '#45B7D1'
}

# Chord diagrams
chord_cmap = 'coolwarm'  # Diverging colormap for relationships

# Radar charts
radar_colors = plt.cm.tab10(np.linspace(0, 1, 10))  # Up to 10 series

Examples

Complete Specialized Visualization Workflow

from metainformant.visualization import specialized as spec_viz
import numpy as np
import pandas as pd

# Create comprehensive specialized visualization suite
fig, axes = plt.subplots(2, 3, figsize=(18, 12))

# Venn diagram
sets = {
    'Group_A': set(range(50)),
    'Group_B': set(range(25, 75)),
    'Group_C': set(range(40, 90))
}
spec_viz.plot_venn_diagram(sets, ax=axes[0,0])

# Chord diagram
matrix = np.random.rand(5, 5)
matrix = (matrix + matrix.T) / 2
np.fill_diagonal(matrix, 0)
labels = ['A', 'B', 'C', 'D', 'E']
spec_viz.plot_chord_diagram(matrix, labels, ax=axes[0,1])

# Radar chart
radar_data = pd.DataFrame({
    'Sample1': [0.8, 0.6, 0.9, 0.7, 0.5],
    'Sample2': [0.6, 0.8, 0.7, 0.6, 0.8],
    'Sample3': [0.7, 0.7, 0.8, 0.9, 0.6]
})
spec_viz.plot_radar_chart(radar_data, ax=axes[0,2])

# Circular bar plot
values = np.random.rand(8) * 100
labels = ['Jan', 'Feb', 'Mar', 'Apr', 'May', 'Jun', 'Jul', 'Aug']
spec_viz.plot_circular_barplot(values, labels, ax=axes[1,0])

# Alluvial diagram (placeholder - requires alluvial package)
# alluvial_data = pd.DataFrame({
#     'Stage1': ['A', 'B', 'C'] * 10,
#     'Stage2': ['X', 'Y', 'Z'] * 10,
#     'Count': np.random.randint(1, 10, 30)
# })
# spec_viz.plot_alluvial_diagram(alluvial_data, ['Stage1', 'Stage2'], ax=axes[1,1])

# Sankey diagram (static version)
flows = [
    ('Input', 'Process_A', 40),
    ('Input', 'Process_B', 60),
    ('Process_A', 'Output_1', 30),
    ('Process_A', 'Output_2', 10),
    ('Process_B', 'Output_2', 40),
    ('Process_B', 'Output_3', 20)
]
spec_viz.plot_sankey_diagram(flows, ax=axes[1,2])

plt.tight_layout()
plt.savefig("specialized_visualizations.png", dpi=300, bbox_inches='tight')

# Create separate interactive plots
# UpSet plot
upset_data = {
    'Set1': {'A', 'B', 'C', 'D'},
    'Set2': {'B', 'C', 'E', 'F'},
    'Set3': {'A', 'D', 'E', 'G'}
}
upset_fig = spec_viz.plot_upset_plot(upset_data)

# Interactive Sankey
interactive_sankey = spec_viz.plot_sankey_diagram(flows, interactive=True)

Troubleshooting

Common Issues

Missing Dependencies: Install optional packages for full functionality
Large Data: Reduce complexity for better performance
Color Conflicts: Use custom color schemes for clarity
Layout Issues: Adjust figure sizes for complex diagrams

Data Requirements

Venn Diagrams: Dictionary of set objects (up to 5 sets recommended)
Chord Diagrams: Square relationship matrix
Sankey Diagrams: List of (source, target, value) tuples
Alluvial Diagrams: Pandas DataFrame with stage columns
Radar Charts: DataFrame with variables as rows, observations as columns

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Specialized Visualization

Overview

Module Functions

Set Analysis

Venn Diagrams

UpSet Plots

Flow Visualization

Sankey Diagrams

Alluvial Diagrams

Network Analysis

Chord Diagrams

Circular Network Layout

Multi-dimensional Data

Radar Charts

Circular Bar Plots

Integration Examples

With Multi-Omics Data

With GWAS Results

With Ontology Analysis

Output Options

Dependencies

Performance Considerations

Color Schemes

Examples

Complete Specialized Visualization Workflow

Troubleshooting

Common Issues

Data Requirements

Related Documentation

FilesExpand file tree

specialized.md

Latest commit

History

specialized.md

File metadata and controls

Specialized Visualization

Overview

Module Functions

Set Analysis

Venn Diagrams

UpSet Plots

Flow Visualization

Sankey Diagrams

Alluvial Diagrams

Network Analysis

Chord Diagrams

Circular Network Layout

Multi-dimensional Data

Radar Charts

Circular Bar Plots

Integration Examples

With Multi-Omics Data

With GWAS Results

With Ontology Analysis

Output Options

Dependencies

Performance Considerations

Color Schemes

Examples

Complete Specialized Visualization Workflow

Troubleshooting

Common Issues

Data Requirements

Related Documentation