Introduction
M-DeepAssembly2 provides both static and multiple conformations prediction services for multi-domain proteins. M-DeepAssembly generates diverse ensembles using a multi-objective protein conformational space sampling algorithm, partially alleviating the challenges posed by weak evolutionary signals and large protein structures. This approach not only offers new perspectives for exploring the functions of multi-domain proteins but also enables multi-conformational studies based on the generated ensembles. Building on this, M-DeepAssembly2 incorporates a flexible residue prediction network to perform directed adjustment of information in distance maps, enabling multiple conformations prediction of proteins. This provides a novel framework for investigating the dynamic behaviors and regulatory mechanisms of molecular machines, signaling proteins, and allosteric proteins.
Input
M-DeepAssembly News
- 2025/08/06: M-DeepAssembly Enhanced Version M-DeepAssembly2 is released.
- 2024/01/24: DeepAssembly Enhanced Version M-DeepAssembly is released.
- 2023/05/03: The server adds protein complex structure assembly functioin.
- 2022/05/28: DeepAssembly was randked No. 1 for the protein structure prediction in the 1 week CAMEO blind test.
- 2022/05/21: We developed DeepAssembly for multi-domain protein structure assembly.
Reference
- Xinyue Cui✝, Yuhao Xia✝, Minghua Hou, Xuanfeng Zhao, Suhui Wang and Guijun Zhang*. M-DeepAssembly:Enhanced DeepAssembly based on multi-objective protein conformation sampling. BMC bioinformatics, 26, 120 (2025). DOI: https://bmcbioinformatics.biomedcentral.com/articles/10.1186/s12859-025-06131-2 .
- Yuhao Xia✝,Kailong Zhao✝,Dong Liu,Xiaogen Zhou,Guijun Zhang*. Multi-domain and complex protein structure prediction using inter-domain interactions from deep learning. Communications biology, 6, 1221 (2023). DOI: https://www.nature.com/articles/s42003-023-05610-7.