Residue Interaction Network (RIN) analysis

This repository provides scripts for performing Residue Interaction Network (RIN) analysis of biomolecular structures.

The workflow involves two main steps:

1. Affinity Matrix Generation using the Fortran script affinity_matrix.f.
2. Hub Residue Identification & Graph Spectral Analysis using the MATLAB script graph_RIN.m.

📖 Overview

1. Affinity Matrix Generation

The affinity_matrix.f script computes the affinity (aᵢⱼ), or local interaction strength, of residue pairs from an input PDB file. aij=Nij/(sqrt(Ni*Nj)), where Nij is the number of residue-residue atom contacts within 4.5 Å; Ni and Nj are the number of heavy atoms in residues i and j.

2. Hub Residue Identification & Graph Spectral Analysis

The graph_RIN.m MATLAB script:

• Represents the protein structure as a residue interaction network (RIN).
  • Nodes: Residue/nucleotide centers of mass.
  • Edges: Weighted by the inverse of local interaction strength, (1/aij).
• Calculates betweenness centrality (CB) for each node.
• Identifies hub residues (residues in the top 5% of CB values).
• Performs graph spectral analysis to partition the network into node clusters.

🛠 Dependencies

• Fortran Compiler (e.g., gfortran) → Required for affinity_matrix.f.
• MATLAB R2019a or later → Required for graph_RIN.m.

🚀 Usage

Step 1: Affinity Matrix Generation (affinity_matrix.f)

Input:

• PDB file containing atomic coordinates (e.g., 7aku-apo.pdb).

Output:

• coordmin1.pdb: Residue centroid coordinates.
• connectivity.out: Local interaction strengths (aij).

Key Parameters:

• rcutt: Distance cutoff for residue interactions (default: 4.5 Å).
• rno: Number of residues in the system.
• nresidue: Number of atoms in the system.

Command-line Usage:

# Compile
gfortran -o affinity_matrix affinity_matrix.f  

# Run
./affinity_matrix

(Ensure the input PDB file is in the same directory.)

Step 2: Hub Residue Identification & Graph Spectral Analysis (graph_RIN.m)

Input:

• connectivity.out (from Step 1).
• coordmin1.pdb (from Step 1).

Output:

• betweenness.out: Betweenness centrality (CB) values.

• top_5_percent.out: Hub residues (top 5% of CB values).

• 10modes.txt: First 10 non-zero eigenvectors of the Laplacian matrix.

• 10modes_normalized.txt: Normalized eigenvectors (-1 or +1).

  Figures:

• Figure 1: Residue interaction network showing nodes colored by CB. The plot can be rotated in 3D.

• Figure 2: CB of the residues.

• Figure 3: CB distributions.

• Figure 4: Spectral analysis shown on the network (Fiedler vector).

MATLAB Usage:

• Open MATLAB R2019a or later

• Ensure that connectivity.out and coordmin1.pdb are present in the same directory

  run('graph_RIN.m')

📂 Example Input & Output

The example_data/ directory provides sample files:

• Input
  • 7aku-apo.pdb (example PDB structure).
  • affinity_matrix.f (Fortran script).
  • graph_RIN.m (MATLAB script).
• Output
  • coordmin1.pdb (centroid coordinates).
  • connectivity.out (affinity values).
  • betweenness.out (betweenness centrality values).
  • top_5_percent.out (hub residues).
  • 10modes.txt & 10modes_normalized.txt (spectral vectors).

📜 License

This project is licensed under the MIT License. See the LICENSE file for details.

📌 Citation

If you use this work in your research, please cite this study and the associated publication:

Kurkcuoglu O. (2018) Exploring allosteric communication in multiple states of the bacterial ribosome using residue network analysis, Turkish Journal of Biology, 42(5): 392–404. DOI: 10.3906/biy-1802-77

✉️ Contact

For questions or issues, please contact:

Dr. Ozge Kurkcuoglu — 📧 olevitas@itu.edu.tr