科研之友
微信
新浪微博
Facebook
分享链接
摘要
The role of a protein depends heavily on its 3D shape, which is composed of semi-independent three-dimensional blocks called domains. Domains fold independently and constitute units of evolution. Most proteins contain multiple domains that are associated with a particular functions; moreover, the same domain can be found in different proteins. Automated recognition of domains can make prediction of proteins function easier and can support the analysis of proteins. Here, we propose a novel algorithm designed for domain recognition by identification of domain boundaries in the protein structure. The proposed algorithm uses a contact graph and an iterative approach to find meaningful clusters corresponding to the protein domains. The distinctive feature of the method is its effective complexity, that improves over other well-known methods, while holding a comparable level of correct domain assignments.
- 出版日期2016-6
