Loop-Diffusion: an equivariant diffusion model for designing and scoring protein loops
Fuente:
arXiv
Guardado en:
| Autores principales: | , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2024
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866909327570239488 |
|---|---|
| author | Borisiak, Kevin Visani, Gian Marco Nourmohammad, Armita |
| author_facet | Borisiak, Kevin Visani, Gian Marco Nourmohammad, Armita |
| contents | Predicting protein functional characteristics from structure remains a central problem in protein science, with broad implications from understanding the mechanisms of disease to designing novel therapeutics. Unfortunately, current machine learning methods are limited by scarce and biased experimental data, and physics-based methods are either too slow to be useful, or too simplified to be accurate. In this work, we present Loop-Diffusion, an energy based diffusion model which leverages a dataset of general protein loops from the entire protein universe to learn an energy function that generalizes to functional prediction tasks. We evaluate Loop-Diffusion's performance on scoring TCR-pMHC interfaces and demonstrate state-of-the-art results in recognizing binding-enhancing mutations. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2409_18201 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Loop-Diffusion: an equivariant diffusion model for designing and scoring protein loops Borisiak, Kevin Visani, Gian Marco Nourmohammad, Armita Biological Physics Machine Learning Quantitative Methods Predicting protein functional characteristics from structure remains a central problem in protein science, with broad implications from understanding the mechanisms of disease to designing novel therapeutics. Unfortunately, current machine learning methods are limited by scarce and biased experimental data, and physics-based methods are either too slow to be useful, or too simplified to be accurate. In this work, we present Loop-Diffusion, an energy based diffusion model which leverages a dataset of general protein loops from the entire protein universe to learn an energy function that generalizes to functional prediction tasks. We evaluate Loop-Diffusion's performance on scoring TCR-pMHC interfaces and demonstrate state-of-the-art results in recognizing binding-enhancing mutations. |
| title | Loop-Diffusion: an equivariant diffusion model for designing and scoring protein loops |
| topic | Biological Physics Machine Learning Quantitative Methods |
| url | https://arxiv.org/abs/2409.18201 |