herbSyner-Finder 0.1.1

Synergistic Ingredient Discovery in Herbal Medicine

HerbSyner_Finder

A General Tool for Synergistic Ingredient Discovery in Herbal Medicine

Overview

HerbSyner_Finder is a computational pipeline designed to discover synergistic ingredient combinations in herbal medicine using network-based approaches and community detection algorithms. This tool helps identify key therapeutic modules and potential synergistic herb combinations for specific diseases through advanced network analysis.

Overview of network framework. (A) Systematic comparison of five prescriptions for CVA on their therapeutic targets. (B) Identify synergistic ingredients of CVA prescriptions through a comprehensive combination landscape (C) Experimental validation of synergistic ingredients.

Features

• Network-based distance calculation: Compute distances between herbs and ingredients in biological networks
• Community detection: Identify communities in herb-ingredient-disease networks using Louvain algorithm
• Key module identification: Detect synergistic modules with main therapeutic effects
• ADMET property filtering: Filter ingredients based on drug-likeness properties
• Multi-disease support: Analyze multiple disease targets simultaneously
• Visualization tools: Generate comprehensive synergy landscapes

Installation

Prerequisites

• Python 3.10.11
• Required Python packages:

pip install communities==2.1.1
pip install networkx==2.8.4
pip install scikit-learn==1.2.2
pip install pandas numpy matplotlib seaborn openpyxl

Install from PyPI

pip install herbSyner_Finder
Install from Source

Clone the repository

git clone https://github.com/[username]/herbSyner_Finder.git
cd herbSyner_Finder

Install the package

pip install -e .

Quick Start

1. Import Required Modules

import pickle
import networkx as nx
import pandas as pd
import herbSyner_Finder
from herbSyner_Finder import Herb_Comb_Community
from herbSyner_Finder import Detection_key_Module_multiple
from herbSyner_Finder import herb_synergy_landscape
from herbSyner_Finder import generate_objects

2. Load Data Objects

g_obj, ingredients_obj, herb_obj, herb_info, fangji, disease_obj, herb_distance_obj = generate_objects.load_obj_data()

3. Set Save Path and File Paths

save_path = "result"
HIf = 'result/herb_ingredient_pairs_pd.csv'
HHf = 'result/herb_herb_dis_pd.csv'
IIf = 'result/ingre_ingre_dis_pd.csv'
DHf = 'result/herb_disease_pd.csv'
DIf = 'result/herb_disease_ingre_pd.csv'
disease_name = "Cough Variant Asthma"

Complete Example Code Here's a complete example showing the full workflow:

import
 pickle
import networkx as nx
import pandas as pd
import herbSyner_Finder
from herbSyner_Finder import Herb_Comb_Community
from herbSyner_Finder import Detection_key_Module_multiple
from herbSyner_Finder import herb_synergy_landscape
from herbSyner_Finder import generate_objects

# Load data objects
g_obj, ingredients_obj, herb_obj, herb_info, fangji, disease_obj, herb_distance_obj = generate_objects.load_obj_data()

# Set save path
save_path = "result"

# Define file paths
HIf = 'result/herb_ingredient_pairs_pd.csv'
HHf = 'result/herb_herb_dis_pd.csv'
IIf = 'result/ingre_ingre_dis_pd.csv'
DHf = 'result/herb_disease_pd.csv'
DIf = 'result/herb_disease_ingre_pd.csv'
disease_name = "Cough Variant Asthma"

# Step 1: Calculate network distance
disease_list = ['Cough Variant Asthma']

# Use herb pair 'MA HUANG','ZI SU YE' as example
combination_herb_list=['MA HUANG','ZI SU YE']

herb_list, herb_herb_dis, ingre_ingre_dis, herb_ingre_pairs = herb_synergy_landscape.cal_combination_distance(herb_info,combination_herb_list)

# Here we use random time = 1 as example to save time
herb_disease_pd, herb_disease_ingre_pd = herb_synergy_landscape.cal_combination_disease(herb_list, herb_distance_obj, disease_list, random_time=1)

# Save distance data from example as files
herb_herb_dis.to_csv('result/herb_herb_distance_example.csv', index=None) 
ingre_ingre_dis.to_csv('result/ingre_ingre_distance_example.csv', index=None)
herb_ingre_pairs.to_csv('result/herb_ingredient_pairs_example.csv', index=None)
herb_disease_pd.to_csv('result/herb_disease_pd_example.csv', index=None)
herb_disease_ingre_pd.to_csv('result/herb_disease_ingre_pd_example.csv', index=None)

# Step 2: Calculate the communities for the complex landscape between herbs
# Reload data generated by step 1
HH = pd.read_csv(HHf).drop_duplicates()
II = pd.read_csv(IIf).drop_duplicates()
DH = pd.read_csv(DHf).drop_duplicates()
DI = pd.read_csv(DIf).drop_duplicates()
ADMETf = 'source_data/ingredient_ADMET_Properties.xlsx'
ADMET = pd.read_excel(ADMETf).drop_duplicates()

# Detect herb communities
herb_commuties_case_obj = Herb_Comb_Community(HH, II, DH, DI, disease_name)
G, communities, DHI_result, subset_data_I, subset_data_H, node_detect_df = herb_commuties_case_obj.get_herb_communities(ADMET_filter = True)

Essential Data Files:

Herb-Ingredient Pairs (herb_ingredient_pairs_pd.csv): Mapping between herbs and their chemical ingredients

Herb-Herb Distances (herb_herb_dis_pd.csv): Network distances between herb pairs

Ingredient-Ingredient Distances (ingre_ingre_dis_pd.csv): Network distances between ingredient pairs

Herb-Disease Associations (herb_disease_pd.csv): Herb-disease therapeutic relationships

Herb-Disease-Ingredient Associations (herb_disease_ingre_pd.csv): Detailed associations between herbs, ingredients, and diseases

Dependencies

communities 2.1.1: Community detection algorithms (GitHub)

networkx 2.8.4: Network analysis and graph operations

scikit-learn 1.2.2: Machine learning utilities

pandas: Data manipulation and analysis

numpy: Numerical computations

matplotlib/seaborn: Visualization

jupyter: For running example notebooks

tqdm: For progress bars in lengthy computations

References

Communities package documentation

NetworkX documentation: https://networkx.org/

Scikit-learn documentation: https://scikit-learn.org/