DRARDT - Drug Repurposing Assessment for Rare Disease Targets

✨ Features

• 🔬 Comprehensive target analysis from multiple databases (UniProt, PubMed, STRING-DB, KEGG, PDB, AlphaFold)
• 💊 DRARDT scoring system (0-3) indicating drug repurposing potential
• 🧬 Optional mutation analysis with 3D definition across PDB and AlphaFold target structures and structural impact prediction using FreeSASA and SimBa-NI method

📁 Core Files

1. drardt.py - Main command-line interface
2. drardt_core.py - DRARDT scoring algorithm
3. drardt_mutations.py - 3D structural analysis (B-factor, pLDDT, RSA, ΔΔG)
4. simba.tsv - Amino acid properties database
5. environment.yml - Conda environment specification

📦 Installation

Requirements

• Conda or Miniconda
• Linux or macOS (recommended)

🎯 Quick Install

# 1. Clone repository
git clone https://github.com/serenafrancisco/drardt.git
cd drardt

# 2. Create environment from file
conda env create -f environment.yml

# Activate environment
conda activate drardt

The environment.yml file automatically installs:

• Python 3.10
• Biopython 1.81
• Requests 2.31.0
• FreeSASA (from conda-forge)

🚀 Usage

Basic Analysis

Analyze a gene target without mutations:

python drardt.py --gene GENE_NAME --email your@email.com

🧬 Analysis with Mutations

Include missense mutation analysis:

python drardt.py --gene GENE_NAME --email your@email.com \
  --mutations R367Q,F409S

📄 Save Results

Save output to a file:

python drardt.py --gene CFTR --email your@email.com --output results.txt

⚙️ Parameters

• --gene (required): Gene name (e.g., BRCA1, TP53, CFTR)
• --email (required): Valid email address (required by NCBI Entrez API)
• --mutations (optional): Comma-separated missense mutation(s) in format A123B (wildtype-position-mutant). If more than one mutation provided, these must be comma-separated.
• --output (optional): Output file path (default: prints to stdout)
• --help: Print help and exit

🔢 DRARDT Scoring

The tool evaluates 5 parameters to calculate a final DRARDT score:

Individual Parameters

1. 📚 Publications (2000-present)

   • Counts PubMed articles mentioning the gene
   • Score: 1-4 points

2. 🔗 Protein Interactors (STRING-DB)

   • Number of experimentally validated protein-protein interactions
   • Score: 1-4 points

3. 🔀 KEGG Pathways

   • Number of metabolic/signaling pathways the gene is involved in
   • Score: 1-3 points

4. ⚛️ PDB Structures

   • Number of available experimental 3D structures
   • Score: 1-4 points

5. AlphaFold2 Prediction

   • Availability of AI-predicted structure
   • Score: 1-2 points

⭐ Final DRARDT Score

Score	Label	Interpretation
0	Very Low	Minimal drug repurposing potential
1	Low	Limited drug repurposing potential
2	High	Good drug repurposing potential
3	Very High	Excellent drug repurposing potential

🧬 Mutation Analysis

When analyzing mutations with a PDB structure, the tool provides:

PDB Coverage

Checks if the mutation position is present in the provided experimental structure.

Relative Solvent Accessibility (RSA)

Calculated using FreeSASA:

• RSA > 20%: Residue is solvent-exposed
• RSA ≤ 20%: Residue is buried in the protein core

ΔΔG Prediction

Stability change prediction using the SimBa-NI method:

• ΔΔG > -1.5 kcal/mol: Mutation is stabilizing or neutral (unlikely to cause unfolding)
• ΔΔG ≤ -1.5 kcal/mol: Mutation is destabilizing (likely to cause protein unfolding)

Formula: ΔΔG = -1.64 + 1.9(RSA/100) + 0.49(Vdiff/100) - 0.12(Hdiff)

Where:

• Vdiff = Volume difference between mutant and wild-type amino acids
• Hdiff = Polarity difference between mutant and wild-type amino acids

Example Output

======================================================================
DRARDT Analysis for AP4M1
======================================================================

GENE INFORMATION
----------------------------------------------------------------------
UniProt ID: O00189
Protein length: 453 amino acids

Disease associations:
  • Spastic paraplegia 50, autosomal recessive: [description]

DRARDT PARAMETERS
----------------------------------------------------------------------

Publications (2000-present): 55
  Score: 2/4

STRING-DB Interactors: 2
  Interactors: AP4B1, AP4E1
  Score: 2/4

KEGG Pathways: 1
  Score: 2/3

PDB Structures: 2
  • PDB ID: 3L81, Method: X-ray, Resolution: 1.60 A, Chains: A=160-453
  • PDB ID: 4MDR, Method: X-ray, Resolution: 1.85 A, Chains: A=160-453
  Score: 2/4

AlphaFold2 Prediction: https://alphafold.ebi.ac.uk/entry/O00189
  Score: 2/2


DRARDT FINAL SCORE
======================================================================
Score: 1/3
Assessment: Low

Interpretation: This target may have limited potential for drug repurposing.

======================================================================
Mutation Analysis
======================================================================


Fetching PDB structures from RCSB...
  Downloaded PDB structure: 3L81
  Downloaded PDB structure: 4MDR
Found 2 PDB structure(s)

Analyzing PDB structure: 3L81

Analyzing PDB structure: 4MDR

Fetching AlphaFold structure...
  Attempting to download from: https://alphafold.ebi.ac.uk/files/AF-O00189-F1-model_v6.pdb
  Response status code: 200
  Successfully saved to: /tmp/tmpl86yh53d.pdb
AlphaFold structure downloaded

Mutation: R367Q
  PDB Coverage: YES
    • 3L81, Chain A
    • 4MDR, Chain A
  3D Structural Quality: PDB (3L81): Medium | PDB (4MDR): Medium | AlphaFold: High
    • PDB 3L81 B-factor (5Å radius): 32.49 Ų (10 residues, 47 atoms)
    • PDB 4MDR B-factor (5Å radius): 43.20 Ų (7 residues, 44 atoms)
    • AlphaFold pLDDT (5Å radius): 95.83 (6 residues, 46 atoms)
  RSA (Relative Solvent Accessibility):
    • 3L81: 62.47% (Solvent-exposed)
    • 4MDR: 47.94% (Solvent-exposed)
    • AlphaFold: 50.24% (Solvent-exposed)
  ΔΔG (Stability prediction):
    • 3L81: -0.60 kcal/mol → Stable
    • 4MDR: -0.87 kcal/mol → Stable
    • AlphaFold: -0.83 kcal/mol → Stable

Mutation: F409S
  PDB Coverage: YES
    • 3L81, Chain A
    • 4MDR, Chain A
  3D Structural Quality: PDB (3L81): High | PDB (4MDR): High | AlphaFold: High
    • PDB 3L81 B-factor (5Å radius): 23.87 Ų (9 residues, 62 atoms)
    • PDB 4MDR B-factor (5Å radius): 28.93 Ų (8 residues, 56 atoms)
    • AlphaFold pLDDT (5Å radius): 96.55 (7 residues, 55 atoms)
  RSA (Relative Solvent Accessibility):
    • 3L81: 2.53% (Buried)
    • 4MDR: 2.80% (Buried)
    • AlphaFold: 2.28% (Buried)
  ΔΔG (Stability prediction):
    • 3L81: -2.57 kcal/mol → Destabilizing
    • 4MDR: -2.56 kcal/mol → Destabilizing
    • AlphaFold: -2.57 kcal/mol → Destabilizing

🗄️ Data Sources

• UniProt: Protein information, disease associations, length
• PubMed/NCBI: Publication counts
• STRING-DB: Protein-protein interactions
• KEGG: Metabolic and signaling pathways
• RCSB PDB: Experimental 3D structures
• AlphaFold: AI-predicted protein structures
• FreeSASA: Solvent accessibility calculations

⚙️ Technical Details

3D Structural Quality Assessment

• Selection method: Any atom within 5Å of mutation CA (Chimera-compatible)
• B-factor thresholds: <30 (High), 30-60 (Medium), ≥60 (Low)
• pLDDT thresholds: >70 (High), 50-70 (Medium), ≤50 (Low)
• Averaging: Per-residue average → average of per-residue averages

RSA Calculation

• Uses FreeSASA with Lee-Richards algorithm
• Normalized to percentage (0-100%)
• Threshold: >20% = exposed, ≤20% = buried

ΔΔG Prediction

• SimBa-NI algorithm (Structure-informed Bayesian model)
• Formula: ΔΔG = -1.64 + 1.9(RSA/100) + 0.49(Vdiff/100) - 0.12(Hdiff)
• Threshold: > -1.5 = Stable, ≤ -1.5 = Destabilizing

📁 Files

drardt/
├── drardt.py              # Main CLI script
├── drardt_core.py         # Core analysis functions
├── drardt_mutations.py    # Mutation analysis functions
├── simba.tsv             # Amino acid properties database
├── environment.yml        # Conda environment specification
└── README.md             # This file

❌ Invalid mutation format

Mutations must be in format A123B:

• A = single letter amino acid code (wild-type)
• 123 = position number
• B = single letter amino acid code (mutant)

Examples: R367Q, F409S, A123V

Importantly, multiple mutations should be comma-separated without spaces (e.g. 'R367Q,F409S').

📧 Contact

• Authors: Serena Francisco, Lorenzo Lamacchia
• Email: serena.francisco@unito.it
• GitHub: @serenafrancisco
• Institution: University of Turin

🙏 Acknowledgments

• Caldararu, Octav et al. “Three Simple Properties Explain Protein Stability Change upon Mutation.” Journal of chemical information and modeling vol. 61,4 (2021): 1981-1988. doi:10.1021/acs.jcim.1c00201
• FreeSASA

📖 Citation

If you use DRARDT in your research, please cite: Francisco, Serena et al. “Restoring adapter protein complex 4 function with small molecules: an in silico approach to spastic paraplegia 50.” Protein science : a publication of the Protein Society vol. 34,1 (2025): e70006. doi:10.1002/pro.70006