Co-expression Network¶
The Co-expression tab visualizes gene correlation networks computed from your normalised count data. Unlike the PPI tab (which requires an external API), this report works for any organism — no internet or STRING database needed.
How It Works¶
Pairwise Pearson correlations are computed between significant DEGs using log2-transformed normalised counts. Gene pairs above the correlation threshold are displayed as a network.
Controls¶
| Control | Range | Default | Description |
|---|---|---|---|
| Max DEGs | 10–200 | 50 | Number of top DEGs (by adjusted p-value) to include |
| Min |correlation| | 0.0–1.0 | 0.70 | Minimum absolute Pearson r to draw an edge |
| Max edges | 10–500 | 100 | Maximum number of edges to display |
Network Graph¶
- Nodes = genes, colored by regulation direction (red = Up, blue = Down, grey = Other; uses sidebar Up/Down/Others colors)
- Node size = scales from the sidebar Dot Size setting, growing with the number of connections (degree)
- Node labels = preferred gene name when available, otherwise gene ID; use the sidebar Font color setting
- Red edges = positive correlation (co-expressed; uses sidebar Up color)
- Blue edges = negative correlation (inversely expressed; uses sidebar Down color)
- Color legend = displayed on the right side of the graph, showing node direction categories (Up-regulated, Down-regulated, Other) and edge correlation types (Positive, Negative). Click a legend entry to toggle visibility.
Data Tables¶
- Hub Genes — Top connected genes ranked by number of connections, with direction, log2FC, and adjusted p-value
- Co-expression Edges — All gene pairs with their Pearson correlation values
- Correlation Matrix — Heatmap of pairwise correlations among connected genes
Tip
Lower the "Min |correlation|" slider to discover weaker but potentially meaningful co-expression patterns. Raise it to focus on the strongest relationships.