InClusterSimilarity.py

import pandas as pd
from rdkit import DataStructs
from rdkit.Chem import AllChem
import numpy as np

# Load the data from CSV
input_csv = 'filtered_compounds_spectral_clustering_with_scaffolds.csv'  # Replace with your actual file path
data = pd.read_csv(input_csv)

# Define the prefix for ECFP4 columns
ecfp_prefix = 'ECFP4_'

# Extract ECFP4 columns
ecfp_cols = [col for col in data.columns if col.startswith(ecfp_prefix)]
if not ecfp_cols:
    raise ValueError("No ECFP4 columns found with the specified prefix.")

# Convert ECFP4 bits to RDKit Morgan Fingerprint format
def create_morgan_fingerprints(df, cols):
    fingerprints = []
    for _, row in df.iterrows():
        bits = row[cols].values
        # Create a list of bits for the fingerprint
        bit_vector = [i for i, bit in enumerate(bits) if bit]
        # Create an RDKit molecule
        mol = AllChem.MolFromSmiles(row['Smiles'])
        if mol is None:
            raise ValueError(f"Invalid SMILES string: {row['Smiles']}")
        # Generate Morgan fingerprint
        morgan_fingerprint = AllChem.GetMorganFingerprintAsBitVect(mol, radius=2, nBits=len(bits))
        fingerprints.append(morgan_fingerprint)
    return fingerprints

# Prepare results container
similarity_results = []

# Process each cluster
for cluster_id in data['Spectral_Cluster'].unique():
    cluster_data = data[data['Spectral_Cluster'] == cluster_id]
    fingerprints = create_morgan_fingerprints(cluster_data, ecfp_cols)
    
    # Calculate pairwise similarities within the cluster
    num_compounds = len(fingerprints)
    for i in range(num_compounds):
        for j in range(i + 1, num_compounds):  # Avoid redundant calculations
            similarity = DataStructs.FingerprintSimilarity(fingerprints[i], fingerprints[j])
            similarity_results.append({
                'Spectral_Cluster': cluster_id,
                'Compound1_ChEMBLID': cluster_data.iloc[i]['ChEMBLID'],
                'Compound2_ChEMBLID': cluster_data.iloc[j]['ChEMBLID'],
                'Similarity': similarity
            })

# Convert results to DataFrame
similarity_df = pd.DataFrame(similarity_results)

# Save to CSV
similarity_csv = 'Spectralcluster_compound_similarity.csv'
similarity_df.to_csv(similarity_csv, index=False)
print(f"Similarity results saved to {similarity_csv}")

# Optional: Print a sample of the results
print(similarity_df.head())