omicverse.metabol.corr_network_plot

omicverse.metabol.corr_network_plot#

omicverse.metabol.corr_network_plot(edges_df, *, ax=None, figsize=(6.5, 5.5), layout='spring', node_size=70, node_color='white', node_edge_color='#34495e', edge_width_scale=2.5, r_column='r', seed=0, with_labels=True, label_font_size=7)[source]#

Draw an edge DataFrame as a NetworkX spring-layout plot.

Renders the output of omicverse.metabol.corr_network() as a correlation graph: nodes are metabolites, edges are pairs whose Spearman / Pearson |r| exceeded the network’s threshold. Edge colour encodes the sign of the correlation (red = positive, blue = negative); edge width is proportional to |r|.

Parameters:

  • edges_df (pd.DataFrame) – Output of corr_network() — needs columns feature_a, feature_b and r (or whatever r_column points at).

  • layout ({'spring', 'circular', 'kamada_kawai'}, default 'spring') – NetworkX layout algorithm.

  • node_size (int (default: 70)) – Standard styling. White-fill / dark-edge nodes read better in publications than filled ones because labels overlay legibly.

  • node_color (str (default: 'white')) – Standard styling. White-fill / dark-edge nodes read better in publications than filled ones because labels overlay legibly.

  • node_edge_color (str (default: '#34495e')) – Standard styling. White-fill / dark-edge nodes read better in publications than filled ones because labels overlay legibly.

  • edge_width_scale (float) – Multiplier on |r| for edge width (so |r|=1 is the thickest edge in the figure).

  • r_column (str, default 'r') – Column to read for edge correlations. Switch to 'r_a' / 'r_b' if you’re plotting one half of a DGCA result.

  • with_labels (bool, default True) – Toggle metabolite labels.

  • label_font_size (int, default 7) – Label size — reduce for dense graphs.

  • seed (int) – RNG seed for the spring-layout (only matters when layout='spring').

Returns:

  • (fig, ax) tuple. If edges_df is empty, draws a centred

  • ”no edges” placeholder and returns the empty axes.

Contents