Interpretation of solution scattering data by combining Bayesian inference with molecular dynamics simulations

使用X射线或中子的小角散射(SAS)是探测蛋白质在溶液中的结构和系综的一种流行方法。然而，对数据的解释仍然具有挑战性，主要是因为信息含量低，以及实验信号中的系统误差。因此，在根据SAXS数据对结构模型进行拟合时，模型经常被过度拟合，模型的不确定性无法量化，或者通常不清楚替代模型是否可以同样地解释实验数据。我们的目标是通过将分子动力学(MD)模拟与严格的统计范式相结合，即结合贝叶斯推理和最大熵原理来克服这些不令人满意的情况。我们的目标是量化与SAS数据匹配的结构模型的模糊性和可信区间，我们的目标是得出无序蛋白质的统计基础解决方案集成。在这一资助期间，我们还将把新的高精度SAS数据与MD模拟结合起来，仔细研究生物分子水合层的结构。

Small-angle scattering (SAS), using either X-rays or Neutrons, is a popular method for probing structures and ensembles of proteins in solution. However, the interpretation of the data has remained challenging, mainly due to the low information content and due to systematic errors in experimental signals. Consequently, upon fitting of structural models against SAXS data, the models are frequently overfitted, the uncertainty of the models cannot be quantified, or it remains often unclear if alternative models may equally explain the experimental data. We aim to overcome these problems by unsatisfactory situation by combining molecular dynamics (MD) simulations with rigorous statistical paradigms, namely with Bayesian inference and with the maximum entropy principle. We aim to quantify the ambiguity and the confidence intervals for structural models fitted against SAS data, and we aim to derive statistically founded solution ensembles of disordered proteins. In this funding period, we will in addition combine novel high-precision SAS data with MD simulations to scrutinize the structure of the hydration layer of biomolecules.