| | --- |
| | license: gpl-3.0 |
| | tags: |
| | - biology |
| | --- |
| | |
| | DiffModeler is a computational tool using a diffusion model to automatically build full protein complex structure from cryo-EM maps at intermediate and low resolution. |
| |
|
| | Copyright (C) 2023 Xiao Wang, Han Zhu, Genki Terashi, Daisuke Kihara, and Purdue University. |
| |
|
| | License: GPL v3. (If you are interested in a different license, for example, for commercial use, please contact us.) |
| |
|
| | Contact: Daisuke Kihara (dkihara@purdue.edu) |
| |
|
| | For technical problems or questions, please reach to Xiao Wang (wang3702@purdue.edu). |
| |
|
| | ## Citation: |
| |
|
| | Xiao Wang, Han Zhu, Genki Terashi & Daisuke Kihara. Protein Complex Structure Modeling with Diffusion Model and AlphaFold in cryo-EM maps.bioArxiv, 2023. |
| | ``` |
| | @article{wang2023DiffModeler, |
| | title={Protein Complex Structure Modeling with Diffusion Model and AlphaFold in cryo-EM maps}, |
| | author={Xiao Wang, Han Zhu, Genki Terashi, and Daisuke Kihara}, |
| | journal={bioArxiv}, |
| | year={2023} |
| | } |
| | ``` |
| |
|
| | ## Free Online Server: https://em.kiharalab.org/algorithm/DiffModeler |
| |
|
| | ## Introduction |
| | Cryogenic electron microscopy (cryo-EM) has been widely employed in experimental settings to determine multi-chain protein complexes, but modeling accuracy greatly diminishes when resolution decreases. At intermediate resolutions of 5-10 Å, even template-based structure fitting presents significant challenges. To tackle this issue, we introduce DiffModeler, a fully automated protein complex structure modeling method that leverages a diffusion model for backbone tracing and structure fitting with AlphaFold predicted single-chain structure. In extensive testing on cryo-EM maps at intermediate resolution, DiffModeler showcased remarkably accurate structure modeling, surpassing existing methods significantly. Notably, we successfully modeled a protein complex consisting of 47 chains, comprising 13,462 residues, with an impressive TM-Score of 0.9. We also further benchmarked DiffModeler for maps at low resolution of 10-20 Å and validated its generalizability with plausible performances. |
| |
|
| | ## Overall Protocol |
| | 1) Backbone tracing from cryo-EM maps at intermediate resolution via diffusion model. |
| | 2) Single-chain structure prediction by AlphaFold. |
| | 3) Single-chain structure fitting using VESPER. |
| | 4) Protein complex modeling by assembling algorithms. |
| |
|
| | ## Github Repo: TBA |
