Correlation Network API

Correlation_network FUNCTION

This API generates a correlation network based on completeness values from KEGGaNOG multi-sample output. Correlation coefficients are calculated for each pathway pair across samples, and a network graph is built using significant correlations above a threshold.

The stronger the correlation the bolder the line

correlation_network(df, threshold=0.5, ...)

PARAMETERS

PARAMETER	DESCRIPTION
df	Input data as a pandas DataFrame TYPE: pd.DataFrame
figsize	Size of the figure (width, height) TYPE: Tuple[int, int] DEFAULT: (12, 6)
threshold	Minimum correlation value to include edge in the network TYPE: float DEFAULT: 0.5
node_size	Size of the nodes TYPE: float DEFAULT: 700.0
node_color	Fill color of nodes TYPE: str DEFAULT: "#A3D5FF"
node_edgecolors	Edge color of nodes TYPE: str DEFAULT: "#03045E"
node_linewidths	Width of node borders TYPE: float DEFAULT: 1.5
label_fontsize	Font size of node labels TYPE: float DEFAULT: 8.0
label_color	Font color of labels TYPE: str DEFAULT: "#03045E"
label_verticalalignment	Vertical alignment of labels TYPE: str DEFAULT: "center"
label_horizontalalignment	Horizontal alignment of labels TYPE: str DEFAULT: "center"
label_weight	Font weight of labels TYPE: str DEFAULT: "normal"
edge_cmap	Colormap for edge correlation strength TYPE: matplotlib.colors.Colormap DEFAULT: plt.cm.coolwarm
cbar_size	Width of the colorbar TYPE: float DEFAULT: 0.5
title	Plot title TYPE: str or None DEFAULT: None
title_fontsize	Font size of title TYPE: float DEFAULT: 16.0
title_color	Font color of title TYPE: str DEFAULT: "black"
title_weight	Font weight of title TYPE: str DEFAULT: "normal"
title_style	Font style of title TYPE: str DEFAULT: "normal"
background_color	Background color of the figure TYPE: str DEFAULT: "white"
save_matrix	Optional file path to save the correlation matrix as CSV TYPE: str or None DEFAULT: None

---

RETURNS

Returns an object with:

• fig: the Matplotlib figure object
• ax: the Matplotlib axis object

The object includes .plotfig() and .savefig() methods for convenience.

---

USAGE EXAMPLE

First, run KEGGaNOG in multi-sample mode:

Where to get demo data? Here

! KEGGaNOG -M -i demo_data/listFile.txt -o KEGGaNOG_multi_output

Then use the correlation network plot function:

import kegganog as kgn
import pandas as pd

df = pd.read_csv("KEGGaNOG_multi_output/merged_pathways.tsv", sep="\t")

kgnnet = kgn.correlation_network(
    df,
    node_size=780,
    label_fontsize=7,
    figsize=(8,5)
)

# To show the plot:
kgnnet.plotfig()

Sample output:

To save the figure:

kgnnet.savefig("corrnet_plot.png", dpi=600)

To export the correlation matrix as CSV:

kgn.correlation_network(df, save_matrix="correlation_matrix.csv")