Skip to content

WEB_INTERFACE.md

Latest commit

 

History

History
281 lines (218 loc) · 6.12 KB

WEB_INTERFACE.md

File metadata and controls

281 lines (218 loc) · 6.12 KB

BBB Permeability Web Interface

Beautiful, interactive web application for predicting blood-brain barrier permeability of molecules.

Features

🎨 Beautiful UI

  • Modern gradient design
  • Responsive layout
  • Interactive visualizations
  • Real-time predictions

📊 Comprehensive Analysis

  • BBB Permeability Score (0-1 scale)
  • Category Classification (BBB+, BBB±, BBB-)
  • Molecular Properties (MW, LogP, TPSA, etc.)
  • Drug-likeness Metrics
  • BBB Rule Compliance
  • Warning System for suboptimal properties

🔬 Input Methods

  1. Common Molecules - Select from 20+ pre-loaded molecules

    • CNS Drugs (Caffeine, Cocaine, Morphine, etc.)
    • Simple Molecules (Ethanol, Benzene, Glucose)
    • Amino Acids (Glycine, Alanine, Tryptophan)
    • Neurotransmitters (Dopamine, Serotonin, GABA)

  2. SMILES String - Direct SMILES input for any molecule

  3. Molecule Name (Beta) - Type common drug names

📈 Visualizations

  • Gauge Chart - BBB score visualization
  • Radar Chart - Drug-likeness profile
  • Bar Chart - Molecular properties
  • Color-coded Results - Instant visual feedback

💾 Export Options

  • CSV export for spreadsheet analysis
  • JSON export for programmatic use

Installation

# Install required packages
pip install streamlit plotly

# Or install all requirements
pip install -r requirements.txt

Usage

Launch the Web Interface

streamlit run app.py

Or with environment variable for OpenMP:

# Windows
set KMP_DUPLICATE_LIB_OK=TRUE
streamlit run app.py

# Linux/Mac
export KMP_DUPLICATE_LIB_OK=TRUE
streamlit run app.py

The app will open in your default browser at http://localhost:8501

Quick Start Guide

  1. Select Input Mode in the sidebar

    • Choose "Common Molecules" for quick testing
    • Choose "SMILES String" for custom molecules

  2. Select or Enter Molecule

    • Browse categories (CNS Drugs, Amino Acids, etc.)
    • Or paste a SMILES string

  3. Click "Predict BBB Permeability"

    • Get instant results with visualizations

  4. Analyze Results

    • View BBB score and category
    • Check molecular properties
    • Review warnings if any

  5. Export Results (optional)

    • Download as CSV or JSON

Interface Sections

Sidebar

  • Input Mode Selection
  • Model Information (MAE, parameters, architecture)
  • Category Guide (BBB+, BBB±, BBB-)
  • About Section

Main Panel

  • Input Section - Select/enter molecules
  • Prediction Button - Trigger analysis
  • Results Display:
    • Color-coded category box
    • BBB score gauge
    • Drug-likeness radar
    • Property metrics
    • Detailed analysis
    • Warning system
    • Export buttons

Examples

Example 1: CNS Drug (Caffeine)

Category: BBB+ (High permeability)
Score: 0.782
MW: 194.2 Da
LogP: -1.03
TPSA: 61.8 A^2

Example 2: Amino Acid (Glycine)

Category: BBB- (Low permeability)
Score: 0.114
MW: 75.1 Da
LogP: -0.97
TPSA: 63.3 A^2

Example 3: Aromatic (Benzene)

Category: BBB+ (High permeability)
Score: 0.802
MW: 78.1 Da
LogP: 1.69
TPSA: 0.0 A^2

Common Molecules Database

The app includes 20+ common molecules:

CNS Drugs:

  • Caffeine, Cocaine, Morphine, Nicotine
  • Aspirin, Ibuprofen, Acetaminophen
  • Propranolol

Simple Molecules:

  • Ethanol, Benzene, Toluene, Glucose

Amino Acids:

  • Glycine, Alanine, Tryptophan

Neurotransmitters:

  • Dopamine, Serotonin, GABA

Technical Details

Model

  • Architecture: Hybrid GAT+GraphSAGE GNN
  • Parameters: 649,345
  • Validation MAE: 0.0967
  • Training Dataset: 42 curated compounds

Visualizations

  • Gauge Chart: Real-time BBB score with thresholds
  • Radar Chart: Drug-likeness across 5 properties
  • Bar Chart: Comprehensive molecular properties

Color Scheme

  • Green: BBB+ (High permeability, ≥0.6)
  • Orange: BBB± (Moderate permeability, 0.4-0.6)
  • Red: BBB- (Low permeability, <0.4)

Troubleshooting

Model Not Found

Error: Failed to load model

Solution: Train the model first:

python train_gnn.py

OpenMP Error

OMP: Error #15: Initializing libiomp5md.dll

Solution: Set environment variable:

set KMP_DUPLICATE_LIB_OK=TRUE  # Windows
export KMP_DUPLICATE_LIB_OK=TRUE  # Linux/Mac

Port Already in Use

Error: Port 8501 is already in use

Solution: Specify a different port:

streamlit run app.py --server.port 8502

Customization

Add More Molecules

Edit COMMON_MOLECULES dictionary in app.py:

COMMON_MOLECULES = {
    "Your Molecule": "SMILES_STRING",
    # Add more here
}

Change Theme

Create .streamlit/config.toml:

[theme]
primaryColor = "#667eea"
backgroundColor = "#ffffff"
secondaryBackgroundColor = "#f0f2f6"
textColor = "#262730"
font = "sans serif"

Modify Visualizations

Edit the chart creation functions in app.py:

  • create_gauge_chart() - BBB score gauge
  • create_property_radar() - Drug-likeness radar
  • create_property_bars() - Property bars

Performance

  • Prediction Time: <1 second per molecule
  • Batch Processing: Supported via API mode
  • Concurrent Users: Streamlit caching enables multi-user support

Future Enhancements

Planned features:

  • Molecule drawing interface (JSME/RDKit)
  • Batch upload (CSV/Excel)
  • 3D molecule visualization
  • Historical predictions tracking
  • Comparison mode (multiple molecules)
  • API endpoint mode
  • Mobile-optimized view
  • Dark theme support

Screenshots

The interface includes:

  1. Header - Beautiful gradient title
  2. Sidebar - Settings and information
  3. Input Section - Multiple input modes
  4. Results Panel - Comprehensive analysis
  5. Visualizations - Interactive charts
  6. Export Options - Download results

Support

For issues or questions:

  • Check README.md for system documentation
  • Review RESULTS.md for model performance
  • See example predictions in demo.py

License

Part of the BBB Permeability Prediction System.

Launch the app: streamlit run app.py

Enjoy predicting BBB permeability with beautiful visualizations! 🧬✨