Update app.py

app.py

@@ -898,6 +898,50 @@ def main():
 - **Artificial Superintelligence (ASI)**  
   - Not yet achieved  
 
+
+# 🧬 Innovative Architecture of AlphaFold2: A Hybrid System
+
+## 1. 🔢 Input Sequence  
+- The process starts with an **input sequence** (protein sequence).  
+
+## 2. 🗄️ Database Searches  
+- **Genetic database search** 🔍  
+  - Searches genetic databases to retrieve related sequences.  
+- **Structure database search** 🔍  
+  - Searches structural databases for template structures.  
+- **Pairing** 🤝  
+  - Aligns sequences and structures for further analysis.  
+
+## 3. 🧩 MSA (Multiple Sequence Alignment)  
+- **MSA representation** 📊 (r,c)  
+  - Representation of multiple aligned sequences used as input.  
+
+## 4. 📑 Templates  
+- Template structures are paired to assist the model.  
+
+## 5. 🔄 Evoformer (48 blocks)  
+- A **deep learning module** that refines representations:  
+  - **MSA representation** 🧱  
+  - **Pair representation** 🧱 (r,c)  
+
+## 6. 🧱 Structure Module (8 blocks)  
+- Converts the representations into:  
+  - **Single representation** (r,c)  
+  - **Pair representation** (r,c)  
+
+## 7. 🧬 3D Structure Prediction  
+- The structure module predicts the **3D protein structure**.  
+- **Confidence levels**:  
+  - 🔵 *High confidence*  
+  - 🟠 *Low confidence*  
+
+## 8. ♻️ Recycling (Three Times)  
+- The model **recycles** its output up to three times to refine the prediction.  
+
+## 9. 📚 Reference  
+**Jumper, J., et al. (2021).** Highly Accurate Protein Structure Prediction with AlphaFold. *Nature.*  
+🔗 [Nature Publication Link](https://www.nature.com/articles/s41586-021-03819-2)  
+
     """
     st.sidebar.markdown(markdownPapers)