Life Sciences Division, Advanced Light Source, Lawrence Berkeley National Laboratory, Berkeley, California 94720, USA.
Nature. 2013 Apr 25;496(7446):477-81. doi: 10.1038/nature12070.
Modern small-angle scattering (SAS) experiments with X-rays or neutrons provide a comprehensive, resolution-limited observation of the thermodynamic state. However, methods for evaluating mass and validating SAS-based models and resolution have been inadequate. Here we define the volume of correlation, Vc, a SAS invariant derived from the scattered intensities that is specific to the structural state of the particle, but independent of concentration and the requirements of a compact, folded particle. We show that Vc defines a ratio, QR, that determines the molecular mass of proteins or RNA ranging from 10 to 1,000 kilodaltons. Furthermore, we propose a statistically robust method for assessing model-data agreements (χ(2)free) akin to cross-validation. Our approach prevents over-fitting of the SAS data and can be used with a newly defined metric, RSAS, for quantitative evaluation of resolution. Together, these metrics (Vc, QR, χ(2)free and RSAS) provide analytical tools for unbiased and accurate macromolecular structural characterizations in solution.
现代小角散射(SAS)实验,无论是 X 射线还是中子,都提供了对热力学状态的全面、分辨率受限的观察。然而,评估质量以及验证基于 SAS 的模型和分辨率的方法一直不够完善。在这里,我们定义了关联体积 Vc,这是一个从散射强度中导出的 SAS 不变量,它与粒子的结构状态有关,但与浓度以及对紧凑、折叠粒子的要求无关。我们表明,Vc 定义了一个比值 QR,它可以确定 10 到 1000 千道尔顿范围内的蛋白质或 RNA 的分子量。此外,我们提出了一种用于评估模型-数据一致性(χ(2)free)的统计稳健方法,类似于交叉验证。我们的方法可以防止 SAS 数据的过度拟合,并且可以与新定义的度量标准 RSAS 一起用于定量评估分辨率。这些度量标准(Vc、QR、χ(2)free 和 RSAS)共同为溶液中无偏且准确的大分子结构特征提供了分析工具。