Amyloid databases: mapping the aggregation universe 🧬📊

A comprehensive review of amyloid-related databases, highlighting how experimental data and bioinformatics resources drive advances in protein aggregation research.

Keywords

amyloids, protein aggregation, databases, bioinformatics, amyloid prediction, protein misfolding, aggregation

---

📌 Project highlights

• 🧬 Comprehensive overview of amyloid & aggregation databases
  
• 📊 Covers sequence, structural and interaction resources
  
• 🔗 Highlights connections between databases and prediction tools
  
• ⚠️ Identifies key limitations in current resources
  
• 🚀 Provides curated list of databases: link

---

🎉 New review out! This one is less about a single tool and more about the entire ecosystem of amyloid data 😄

🔗 Explore the resources

• 🌐 Full database list
  
• 📚 Paper (open access)

👉 This is basically a map of the amyloid bioinformatics landscape.

---

🎧 Audio summary

Too many databases to remember? Same 😄

👉 We’ve added a short audio overview 🎧 so you don’t have to memorize all of them.

🔬 What is this about?

Amyloid aggregation is a complex, multi-factorial process involving:

• sequence features
  
• 3D structure
  
• environmental conditions

👉 and it underlies:

• neurodegenerative diseases
  
• biotechnological challenges
  
• functional biological processes

Because of this complexity:

👉 researchers have built many specialized databases to organize experimental knowledge

---

🧠 What we reviewed

We systematically analyzed amyloid-related databases, grouping them into:

🧬 Sequence-based databases

• focus on aggregation-prone regions (APRs)
  
• example: AmyLoad, AmyPro

🧊 Structure-based databases

• store 3D fibril structures
  
• example: Amyloid Atlas

🔗 Interaction databases

• capture cross-interactions between amyloids
  
• example: AmyloGraph

👉 Each database captures different aspects of aggregation.

---

📊 Key insight: fragmentation problem

There is no single “perfect” database.

Instead:

• each resource focuses on a specific niche
  
• data formats and annotations differ
  
• integration is difficult

👉 Result:

❌ no unified benchmark dataset
❌ hard to compare prediction tools
❌ fragmented knowledge

---

⚙️ Databases ↔︎ prediction tools (the feedback loop)

One of the most important conclusions:

👉 databases and prediction tools co-evolve

• experimental datasets → enable model development
  
• prediction tools → generate new hypotheses
  
• new experiments → expand databases

👉 A continuous feedback loop driving the field forward.

---

🧬 Examples of this interplay

• AmyloGraph → enabled PACT / AmyloComp (cross-interactions)
  
• AmyloBase → contributed to AGGRESCAN
  
• Waltz datasets → led to WALTZ algorithm

👉 Data → model → better data → better model

---

⚠️ Key limitations (important!)

Across databases:

• 🔍 limited search & filtering
  
• 📤 poor export options
  
• 🧾 incomplete metadata
  
• 🤖 reliance on predictions (with biases)

👉 And most importantly: aggregation is not only sequence-dependent

Environmental factors matter:

• pH
  
• temperature
  
• concentration
  
• cofactors

---

🚀 Why this matters

This review shows:

👉 we have a lot of data
👉 but not yet fully integrated knowledge

Future directions:

• better standardization (e.g. MIRRAGGE)
  
• integration of datasets
  
• ML models using multi-dimensional data