期刊
JOURNAL OF BIOMOLECULAR NMR
卷 51, 期 1-2, 页码 115-122出版社
SPRINGER
DOI: 10.1007/s10858-011-9546-9
关键词
Troponin C; L48Q; Structure prediction; ORBplus; Principal component; Partial least squares
资金
- AHFMR
- Canadian Institutes of Health Research
Laboratories often repeatedly determine the structure of a given protein under a variety of conditions, mutations, modifications, or in a number of states. This approach can be cumbersome and tedious. Given then a database of structures, identifiers, and corresponding H-1,N-15-HSQC NMR spectra for homologous proteins, we investigated whether structural information could be ascertained for a new homolog solely from its H-1,N-15-HSQC NMR spectrum. We addressed this question with two different approaches. First, we used a semi-automated approach with the program, ORBplus. ORBplus looks for patterns in the chemical shifts and correlates these commonalities to the explicit property of interest. ORBplus ranks resonances based on consistency of the magnitude and direction of the chemical shifts within the database, and the chemical shift correlation of the unknown protein with the database. ORBplus visualizes the results by a histogram and a vector diagram, and provides residue specific predictions on structural similarities with the database. The second method we used was partial least squares (PLS), which is a multivariate statistical technique used to correlate response and predictor variables. We investigated the ability of these methods to predict the tertiary structure of the contractile regulatory protein troponin C. Troponin C undergoes a closed-to-open conformational change, which is coupled to its function in muscle. We found that both ORBplus and PLS were able to identify patterns in the H-1,N-15-HSQC NMR data from different states of troponin C that correlated to its conformation.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据