Ulm University, Institute of Biophysics, Albert-Einstein-Allee 11, Ulm 89069, Germany.
J Chem Phys. 2018 Mar 28;148(12):123308. doi: 10.1063/1.5006477.
Modern hybrid structural analysis methods have opened new possibilities to analyze and resolve flexible protein complexes where conventional crystallographic methods have reached their limits. Here, the Fast-Nano-Positioning System (Fast-NPS), a Bayesian parameter estimation-based analysis method and software, is an interesting method since it allows for the localization of unknown fluorescent dye molecules attached to macromolecular complexes based on single-molecule Förster resonance energy transfer (smFRET) measurements. However, the precision, accuracy, and reliability of structural models derived from results based on such complex calculation schemes are oftentimes difficult to evaluate. Therefore, we present two proof-of-principle benchmark studies where we use smFRET data to localize supposedly unknown positions on a DNA as well as on a protein-nucleic acid complex. Since we use complexes where structural information is available, we can compare Fast-NPS localization to the existing structural data. In particular, we compare different dye models and discuss how both accuracy and precision can be optimized.
现代混合结构分析方法为分析和解决传统晶体学方法已达到极限的柔性蛋白质复合物开辟了新的可能性。在这里，Fast-Nano-Positioning System（Fast-NPS）是一种有趣的方法，它是一种基于贝叶斯参数估计的分析方法和软件，允许基于单分子Förster 共振能量转移（smFRET）测量对附着在大分子复合物上的未知荧光染料分子进行定位。然而，基于这种复杂计算方案的结果得出的结构模型的精度、准确性和可靠性往往难以评估。因此，我们提出了两个原理验证基准研究，我们使用 smFRET 数据来定位 DNA 以及蛋白质-核酸复合物上假定的未知位置。由于我们使用的是结构信息可用的复合物，因此我们可以将 Fast-NPS 定位与现有的结构数据进行比较。特别是，我们比较了不同的染料模型，并讨论了如何优化准确性和精度。
