NanoProNet: An “ImageNet” dataset for protein corona research!
📌 2.76 M annotated nanomaterial–protein samples 📌
2.76M+
Nanomaterial–protein
samples
33.7k+
Number of
unique proteins
28
Nanomaterial, incubation
& separation parameters
Features
AI at the forefront: empowering nanomaterial-protein interaction research 🤖
Largest public dataset
Over 2.76 million annotated nanomaterial–protein interaction samples and 33k unique proteins to advance research and model training.
Support generalizable predictions
The use of universal text and protein language models supports generalized prediction on unseen samples and proteins.
Easy to use
Researchers with a basic machine learning background can easily follow the detailed guidelines and clear usage instructions provided.
High predictive performance
The model is capable of accurate base predictions, handling predictions with missing feature information effectively, and generalizing reliably to unseen data.
Adaptable with fine-tuning
Serving as a foundation model, it can be fine-tuned to specific applications, improving its ability to learn from few examples.
Active research community
This will drive progress in protein corona research, positioning it as a vital component in the rapidly evolving field of AI for Science.
From the authors
We are thankful for the support and collaboration that made this work possible.
"We hope this dataset and model accelerate AI-driven protein corona research and its nanomedicine and broader applications!"
"We are committed to application-driven and trustworthy AI research, exploring its broad applications in industry, science, and art."
