|
|
--- |
|
|
license: apache-2.0 |
|
|
pipeline_tag: tabular-regression |
|
|
tags: |
|
|
- biology |
|
|
datasets: |
|
|
- Allanatrix/ProtienBank |
|
|
metrics: |
|
|
- accuracy |
|
|
--- |
|
|
|
|
|
# NexaBio: Advanced Protein Structure Prediction Models |
|
|
|
|
|
**NexaBio** is a sophisticated two-stage model suite designed for high-accuracy protein structure prediction from amino acid sequences. It comprises two complementary models: |
|
|
|
|
|
- **NexaBio_1**: A Convolutional Neural Network (CNN) and Bidirectional LSTM (BiLSTM) model for secondary structure prediction. |
|
|
- **NexaBio_2**: A Variational Autoencoder (VAE) and Diffusion-based model for tertiary (3D) structure prediction. |
|
|
|
|
|
NexaBio is a core component of the [Nexa Scientific Model Suite](https://huggingface.co/spaces/Allanatrix/NexaHub), a collection of machine learning models advancing scientific discovery. |
|
|
|
|
|
## Model Overview |
|
|
|
|
|
### NexaBio_1: Secondary Structure Prediction |
|
|
- **Architecture**: CNN combined with BiLSTM for robust sequence modeling. |
|
|
- **Input**: Amino acid sequence (one-hot encoded or embedded). |
|
|
- **Output**: Secondary structure classifications (e.g., Helix, Sheet, Coil). |
|
|
- **Use Case**: Identification of local structural motifs and protein folding patterns. |
|
|
|
|
|
### NexaBio_2: Tertiary Structure Prediction |
|
|
- **Architecture**: VAE integrated with a Diffusion Model for generative 3D modeling. |
|
|
- **Input**: Amino acid sequence (optionally augmented with secondary structure predictions). |
|
|
- **Output**: 3D coordinates of protein backbone atoms. |
|
|
- **Use Case**: Full tertiary structure prediction for structural analysis and design. |
|
|
|
|
|
## Applications |
|
|
- **Structural Bioinformatics**: Enabling precise protein structure analysis for research. |
|
|
- **Drug Discovery**: Supporting protein-ligand interaction studies and therapeutic design. |
|
|
- **Protein Engineering**: Facilitating the design of novel proteins for industrial and medical applications. |
|
|
- **Synthetic Biology**: Generating protein structures for biotechnological innovation. |
|
|
- **Academic Research**: Serving as a tool for educational and exploratory studies. |
|
|
|
|
|
## Getting Started |
|
|
|
|
|
### Example Usage |
|
|
```python |
|
|
from transformers import AutoModel |
|
|
|
|
|
# Initialize the secondary structure prediction model |
|
|
model_sec = AutoModel.from_pretrained("Allanatrix/NexaBio_1") |
|
|
|
|
|
# Initialize the tertiary structure prediction model |
|
|
model_ter = AutoModel.from_pretrained("Allanatrix/NexaBio_2") |
|
|
|
|
|
# Process an amino acid sequence (refer to model documentation for input formatting) |
|
|
``` |
|
|
|
|
|
For comprehensive instructions, including inference APIs and preprocessing details, consult the individual model cards on Hugging Face. |
|
|
|
|
|
## Citation and License |
|
|
If you utilize NexaBio in your research or applications, please cite this repository and include a link to the [Nexa R&D Space](https://huggingface.co/spaces/Allanatrix/NexaR&D). |
|
|
The models and associated code are licensed under the **Boost Software License 1.1 (BSL-1.1)**. |
|
|
|
|
|
## Part of the Nexa Scientific Ecosystem |
|
|
Discover other components of the Nexa Scientific Stack: |
|
|
- [Nexa Data Studio](https://huggingface.co/spaces/Allanatrix/NexaDataStudio): Data processing and visualization tools. |
|
|
- [Nexa R&D](https://huggingface.co/spaces/Allanatrix/NexaR&D): Research-focused model development environment. |
|
|
- [Nexa Infrastructure](https://huggingface.co/spaces/Allanatrix/NexaInfrastructure): Scalable ML deployment solutions. |
|
|
- [Nexa Hub](https://huggingface.co/spaces/Allanatrix/NexaHub): Central portal for Nexa resources. |
|
|
|
|
|
--- |
|
|
|
|
|
*Developed and maintained by [Allan](https://huggingface.co/Allanatrix), an independent machine learning researcher specializing in scientific AI and infrastructure.* |
