Kong Yifei, Ma Jianpeng
Graduate Program of Structural and Computational Biology and Molecular Biophysics, Baylor College of Medicine, Houston, TX 77030, USA.
J Mol Biol. 2003 Sep 12;332(2):399-413. doi: 10.1016/s0022-2836(03)00859-3.
Here, we report a new computational method, called sheetminer, for mining beta-sheets in the density maps at intermediate resolutions of 6 to 10A. The method employs a multi-step ad hoc morphological analysis of density maps to identify the unique characteristics of beta-sheets. It was tested on density maps from 12 protein crystal structures that were artificially blurred to intermediate resolutions. There are a total of 35 independent beta-sheets with a wide distribution of morphology. The method successfully located 34 of them and missed only one. The method was also applied to an experimental 9A electron cryomicroscopic structure and an 8A X-ray density map. In both cases, the sheet-searching results were found to agree very well with known high-resolution crystal structures. Collectively, these results demonstrate clearly the robustness of sheetminer in locating the regions belonging to beta-sheets in the intermediate-resolution density maps. Furthermore, sheetminer is completely complementary to all other existing computational methods, including helixhunter and threading algorithms. Their combined usage has the potential to significantly enhance the computational modeling capacity for a much more complete interpretation of structural data at intermediate resolutions, from which extraction of functional information would be more effective. This is particularly important in the field of structural genomics, in which the fast screening approach may not always yield crystals that diffract to atomic resolution. An exciting future application of sheetminer is as a valuable tool for revealing the structures of amyloid fibrils that are rich in beta-motifs.
在此，我们报告一种名为SheetMiner的新计算方法，用于在6至10埃的中等分辨率密度图中挖掘β折叠。该方法采用对密度图的多步特殊形态学分析来识别β折叠的独特特征。它在12个蛋白质晶体结构的密度图上进行了测试，这些密度图被人为模糊到中等分辨率。总共有35个独立的β折叠，形态分布广泛。该方法成功定位了其中34个，仅遗漏了一个。该方法还应用于一个实验性的9埃电子冷冻显微镜结构和一个8埃X射线密度图。在这两种情况下，发现β折叠搜索结果与已知的高分辨率晶体结构非常吻合。总体而言，这些结果清楚地证明了SheetMiner在定位中等分辨率密度图中属于β折叠的区域时的稳健性。此外，SheetMiner与所有其他现有计算方法完全互补，包括螺旋搜索器和穿线算法。它们的联合使用有可能显著提高计算建模能力，以便更完整地解释中等分辨率的结构数据，从中提取功能信息将更有效。这在结构基因组学领域尤为重要，在该领域中，快速筛选方法可能并不总是能得到衍射至原子分辨率的晶体。SheetMiner一个令人兴奋的未来应用是作为揭示富含β基序的淀粉样纤维结构的有价值工具。
