Update README.md
Browse files
README.md
CHANGED
|
@@ -2,6 +2,7 @@
|
|
| 2 |
|
| 3 |
# AIDO.StructurePrediction
|
| 4 |
|
|
|
|
| 5 |
|
| 6 |
## Model Description
|
| 7 |
|
|
@@ -14,7 +15,7 @@ better than AF3 on immunology related structure prediction tasks,
|
|
| 14 |
including those involving antibodies, nanobodies, and antibody-antigen as well as nanobody-antigen interactions.
|
| 15 |
We are currently evaluating its capabilities for other structural interactions involving combinations of proteins, DNA, RNA, and ligands.
|
| 16 |
|
| 17 |
-
|
| 18 |
|
| 19 |
|
| 20 |
## Model Details
|
|
@@ -27,7 +28,7 @@ We are currently evaluating its capabilities for other structural interactions i
|
|
| 27 |
- **Training Strategies**: Adopts advanced training methodologies to optimize model performance and efficiency.
|
| 28 |
|
| 29 |
### Model Architecture
|
| 30 |
-
- **Type**:
|
| 31 |
- **Key Components**:
|
| 32 |
- **Pairformer**: Designed to learn complex relationships from both single sequences and multiple sequence alignments.
|
| 33 |
- **Diffusion Module**: Generates multiple conformations of the structure.
|
|
@@ -83,6 +84,14 @@ inaccurate metric values and hinder head-to-head comparisons between different m
|
|
| 83 |
</div>
|
| 84 |
</div>
|
| 85 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 86 |
## Acknowledgements
|
| 87 |
|
| 88 |
We would like to thank the Protenix team for their public training codebase and contributions to the field of
|
|
@@ -93,14 +102,6 @@ Additionally, we thank DeepMind's AlphaFold3 for their outstanding contributions
|
|
| 93 |
If you find our work useful, please also cite [Protenix](https://github.com/bytedance/Protenix),
|
| 94 |
[MMseqs2](https://github.com/soedinglab/MMseqs2), and [AlphaFold3](https://github.com/google-deepmind/alphafold3).
|
| 95 |
|
| 96 |
-
## License and Disclaimer
|
| 97 |
-
|
| 98 |
-
This project uses the default license of ModelGenerator. For commercial use, please contact us.
|
| 99 |
-
As we built the model based on the public [Protenix](https://github.com/bytedance/Protenix) codebase.
|
| 100 |
-
Please ensure to follow Protenix's license as well. .
|
| 101 |
-
|
| 102 |
-
If you discover a potential security issue in this project, or believe you have found a security vulnerability,
|
| 103 |
-
we request that you notify Genbio AI via email. Please do **not** create a public GitHub issue for security concerns.
|
| 104 |
|
| 105 |
|
| 106 |
# Citation
|
|
|
|
| 2 |
|
| 3 |
# AIDO.StructurePrediction
|
| 4 |
|
| 5 |
+
<img src="assets/vis.png" width="1300" />
|
| 6 |
|
| 7 |
## Model Description
|
| 8 |
|
|
|
|
| 15 |
including those involving antibodies, nanobodies, and antibody-antigen as well as nanobody-antigen interactions.
|
| 16 |
We are currently evaluating its capabilities for other structural interactions involving combinations of proteins, DNA, RNA, and ligands.
|
| 17 |
|
| 18 |
+
|
| 19 |
|
| 20 |
|
| 21 |
## Model Details
|
|
|
|
| 28 |
- **Training Strategies**: Adopts advanced training methodologies to optimize model performance and efficiency.
|
| 29 |
|
| 30 |
### Model Architecture
|
| 31 |
+
- **Type**: Pairformer+Diffusion model architecture.
|
| 32 |
- **Key Components**:
|
| 33 |
- **Pairformer**: Designed to learn complex relationships from both single sequences and multiple sequence alignments.
|
| 34 |
- **Diffusion Module**: Generates multiple conformations of the structure.
|
|
|
|
| 84 |
</div>
|
| 85 |
</div>
|
| 86 |
|
| 87 |
+
## License and Disclaimer
|
| 88 |
+
|
| 89 |
+
This project is under the GenBio AI Community License (also ModelGenerator's license). For commercial use, please contact us.
|
| 90 |
+
|
| 91 |
+
If you discover a potential security issue in this project, or believe you have found a security vulnerability,
|
| 92 |
+
we request that you notify Genbio AI via email. Please do **not** create a public GitHub issue for security concerns.
|
| 93 |
+
|
| 94 |
+
|
| 95 |
## Acknowledgements
|
| 96 |
|
| 97 |
We would like to thank the Protenix team for their public training codebase and contributions to the field of
|
|
|
|
| 102 |
If you find our work useful, please also cite [Protenix](https://github.com/bytedance/Protenix),
|
| 103 |
[MMseqs2](https://github.com/soedinglab/MMseqs2), and [AlphaFold3](https://github.com/google-deepmind/alphafold3).
|
| 104 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 105 |
|
| 106 |
|
| 107 |
# Citation
|