🌾 Physics-Informed Machine Learning to Optimize CRISPR Stomatal Engineering for Climate-Resilient Rice in Telangana

Physics-Informed ML & Agricultural Biotechnology Platform
Authors / Collaborators: Department of Computer Science & Engineering, JNTUH & Research Team
Domain: Biophysics + Physics-Informed ML + Agricultural Biotechnology + Climate Adaptation

---

📌 Abstract & Overview

Rising atmospheric temperatures and extreme Vapor Pressure Deficit (VPD) events in Telangana (Warangal, Nizamabad, Karimnagar, Nalgonda, Khammam) cause severe transpirational water stress and yield penalty in Oryza sativa (rice). While CRISPR-Cas9 genome editing targeting promoter regions of stomatal development genes (OsEPF1, OsEPFL9/10, OsSTOMAGEN) allows precise reduction of stomatal density ($N_s$), non-linear trade-offs between water-use efficiency (WUE) and photosynthetic carbon assimilation ($A$) vary dramatically under regional micro-climates.

This system combines a Medlyn biophysical stomatal conductance model ($g_s$) with a Physics-Informed XGBoost Machine Learning architecture trained on 168 digitized experimental measurements from peer-reviewed literature (Caine et al. 2019, Karavolias et al. 2023, Karavolias et al. 2024) and 11 years of NASA POWER daily climate data (2015–2026).

---

🚀 Key Features

1. Decoupled Architecture:
   • FastAPI Python Backend: REST endpoints & Real-time Server-Sent Events (SSE) streaming predictions.
   • Vite React Tailwind Frontend: Modern agricultural technology user portal with Framer Motion animations, Recharts, Plotly 3D rotatable surfaces, and SHAP explainability.
2. Physics-Informed Hybrid ML Model:
   • Outperforms pure physical Medlyn baselines ($R^2 = +0.183$ vs $R^2 = -0.266$, $+44.9$ percentage point gain).
   • 5-Fold Cross Validation ($0.174 \pm 0.078$).
3. Interactive Visual Cards & Source Research Papers:
   • Clickable lightbox views of authentic Oryza sativa SEM 1000x leaf microscopy and Telangana paddy field cultivation.
   • Direct DOI links to foundational literature.

---

🛠️ Installation & Setup

Prerequisites

• Python 3.9+
• Node.js 18+ and npm

1. Backend Setup

# Clone repository
git clone https://github.com/YOUR_USERNAME/crispr-rice-engineering.git
cd crispr-rice-engineering

# Create virtual environment
python3 -m venv .venv
source .venv/bin/activate

# Install dependencies
pip install -r requirements.txt

# Start FastAPI server
uvicorn backend.main:app --port 8000 --reload

2. Frontend Setup

# Install dependencies
cd frontend
npm install

# Start development server
npm run dev

Open http://localhost:5173 in your browser.

---

📊 Repository Structure

paddy_project/
├── backend/                # FastAPI backend application
│   ├── routes/             # REST & SSE streaming endpoints
│   ├── services/           # Model inference & biophysical calculation logic
│   └── main.py             # FastAPI entry point
├── frontend/               # Vite + React + Tailwind frontend application
│   ├── src/                # React components (ControlPanel, MetricCards, etc.)
│   └── public/figures/     # Authentic rice plant & microscopy images
├── data/                   # Master biological & NASA climate datasets
├── scripts/                # Data pipeline & model training scripts
├── outputs/                # Saved model artifacts & figure exports
├── README.md               # GitHub project overview & documentation
└── requirements.txt        # Python dependency requirements

---

📜 Citation & References

• Caine et al. (2019) – Rice plants with reduced stomatal density exhibit improved water-use efficiency and drought tolerance. New Phytologist. DOI: 10.1111/nph.15344
• Karavolias et al. (2023) – Paralogous OsEPFL genes modulate stomatal density in rice. Plant Physiology. DOI: 10.1093/plphys/kiad183
• Karavolias et al. (2024) – Promoter editing of OsSTOMAGEN tunes stomatal density and yield traits in rice. Plant Biotechnology Journal. DOI: 10.1111/pbi.14464