Terashi Genki, Kagaya Yuki, Kihara Daisuke
Department of Biological Sciences, Purdue University, West Lafayette, Indiana 47907, United States.
Graduate School of Information Sciences, Tohoku University, Aramaki Aza, Aoba 6-3-09, Aoba-Ku, Sendai, Miyagi 980-8579, Japan.
J Chem Inf Model. 2020 May 26;60(5):2634-2643. doi: 10.1021/acs.jcim.9b01110. Epub 2020 Mar 30.
For structural interpretation of cryo-electron microscopy (cryo-EM) density maps that contain multiple chains, map segmentation is an important step. If a map is segmented accurately into regions of individual protein components, the structure of each protein can be separately modeled using an existing modeling tool. Here, we developed new software, MAINMASTseg, for segmenting maps with symmetry. MAINMASTseg is an extension of the MAINMAST cryo-EM protein structure modeling tool, which builds protein structures from a graph structure that captures the distribution of salient density points in the map. MAINMASTseg uses this graph and segments the map by considering symmetry corresponding density points in the graph. We tested MAINMASTseg on a data set of 38 experimentally determined EM density maps. MAINMASTseg successfully identified an individual protein unit for the majority of the maps, which was significantly better than two other popular existing methods, Segger and Phenix. The software is made freely available for academic users at http://kiharalab.org/mainmast_seg.
对于包含多条链的冷冻电子显微镜（cryo-EM）密度图的结构解释而言，图谱分割是重要的一步。如果能将图谱准确分割成各个蛋白质组分的区域，那么就可以使用现有的建模工具分别对每个蛋白质的结构进行建模。在此，我们开发了新软件MAINMASTseg，用于分割具有对称性的图谱。MAINMASTseg是MAINMAST冷冻电镜蛋白质结构建模工具的扩展，该工具通过捕获图谱中显著密度点分布的图形结构构建蛋白质结构。MAINMASTseg利用此图形，并通过考虑图形中对应对称的密度点来分割图谱。我们在一组由38个实验测定的电镜密度图数据集上测试了MAINMASTseg。MAINMASTseg成功地为大多数图谱识别出了单个蛋白质单元，这比另外两种现有的常用方法Segger和Phenix要好得多。该软件可供学术用户在http://kiharalab.org/mainmast_seg免费使用。
