Amyloids don’t aggregate alone 🤝🧬 Meet AmyloGraph

AmyloGraph is a curated database of experimentally validated amyloid–amyloid interactions, enabling systematic exploration of cross-seeding and aggregation modulation.

Keywords

amyloids, protein aggregation, cross-seeding, AmyloGraph, database, bioinformatics, neurodegeneration

📌 Project highlights

• 🧬 First database of amyloid–amyloid interactions
  
• 📊 Contains 883 experimentally validated interactions
  
• 📚 Curated from ~200 publications
  
• 🧠 Standardizes cross-seeding, inhibition, co-aggregation
  
• 🚀 Available as web server + R package

---

🎉 New paper out! This one tackles a fundamental gap:

👉 amyloids interacting with each other… but no structured data 😄

👉 AmyloGraph: a comprehensive database of amyloid–amyloid interactions

---

🔗 Try it yourself

• 🌐 Web server
  
• 💻 GitHub / R package

👉 Explore the amyloid interaction network directly

---

🎧 Audio summary

Amyloids don’t just aggregate…
they talk to each other, accelerate, inhibit, cross-seed 😳

👉 Here’s a short audio overview 🎧 explaining what AmyloGraph brings:

🔬 What is this about?

Amyloids are typically studied as individual aggregating proteins

BUT in reality:

👉 they interact with each other during aggregation

These interactions can:

• ⚡ accelerate fibril formation
  
• 🛑 inhibit aggregation
  
• 🔀 create heterogeneous fibrils

👉 and may explain:

• Alzheimer’s + Parkinson’s overlap
  
• prion-like propagation
  
• complex disease mechanisms

---

⚠️ The core problem

Before AmyloGraph:

• data scattered across hundreds of papers
  
• inconsistent terminology
  
• difficult to compare experiments

👉 even contradictory conclusions

There was:

❌ no centralized dataset
❌ no standardized vocabulary
❌ no way to model interactions

---

🧠 What we built

👉 AmyloGraph = curated database of amyloid–amyloid interactions

---

📊 Data scale

• 883 interactions
  
• 46 amyloid proteins
  
• 172 publications

👉 all experimentally validated

---

⚙️ Standardization (this is key)

AmyloGraph introduces 3 descriptors:

1. ⚡ Effect on aggregation speed
   
2. 🤝 Physical binding evidence
   
3. 🔀 Formation of mixed fibrils

👉 turning messy literature into structured, comparable data

---

🔗 Graph-based representation

Data is presented as:

• nodes → amyloid proteins
  
• edges → interactions

👉 forming an amyloid interaction network

---

📊 Multiple views

• 🔗 graph (network exploration)
  
• 📋 table (filter + download)
  
• 🔍 single interaction (full details)

👉 usable both for:

• humans
  
• ML pipelines

---

📊 Key insights

🧬 Amyloids form a network

Aggregation is not isolated:

👉 proteins influence each other’s behavior

---

🔀 Cross-talk is widespread

Interactions include:

• cross-seeding
  
• inhibition
  
• co-aggregation

👉 all encoded in a standardized way

---

⚠️ Context matters

Results depend on:

• pH
  
• concentration
  
• experimental setup

👉 biology is messy (and the database reflects that)

---

🚀 Why this matters

🧠 Disease understanding

Helps explain:

• overlapping neurodegenerative diseases
  
• propagation of aggregation

---

🤖 Machine learning & modeling

Finally enables:

• prediction of cross-interactions
  
• tools like PACT / AmyloComp built on top

---

💊 Therapeutic strategies

If we know:

👉 which amyloids interact

we can:

• block interactions
  
• design inhibitors
  
• target aggregation pathways

---

💚 BioGenies perspective

This is a foundational dataset paper 🧱

👉 not flashy, but extremely powerful