Bayesian-Maximum-Entropy Reweighting of IDP Ensembles Based on NMR Chemical Shifts

Abstract

Bayesian and Maximum Entropy approaches allow for a statistically sound and systematic fitting of experimental and computational data. Unfortunately, assessing the relative confidence in these two types of data remains difficult as several steps add unknown error. Here we propose the use of a validation-set method to determine the balance, and thus the amount of fitting. We apply the method to synthetic NMR chemical shift data of an intrinsically disordered protein. We show that the method gives consistent results even when other methods to assess the amount of fitting cannot be applied. Finally, we also describe how the errors in the chemical shift predictor can lead to an incorrect fitting and how using secondary chemical shifts could alleviate this problem.

Refinement of α-Synuclein Ensembles Against SAXS Data: Comparison of Force Fields and Methods

“…As also previously shown (Robustelli et al, 2018) there is little transient helical structure in these simulations, though with some variation across force fields. Previous analyses suggest that compaction and secondary structure are only weakly coupled in disordered proteins (Piana et al, 2012;Crehuet et al, 2019;Zerze et al, 2019), and indeed we in general find that reweighting against the SAXS data only has a modest effect on the secondary structure. The M&M simulations, however, do not follow this pattern, but we note here that in contrast to the other simulations, these are two independent simulations.…”

Fitting side-chain NMR relaxation data using molecular simulations

“…In these equations the parameter sets the balance between fitting the data (minimizing 2 ) and keeping as much as possible of the original simulation (maximizing ). We refer the reader to previous literature about the methods overall and how best to select this parameter (Andrae et al, 2010;Hummer and Köfinger, 2015;Cesari et al, 2018;Köfinger et al, 2019;Crehuet et al, 2019;Chen et al, 2019;Orioli et al, 2020).…”

Comparison of Different Reweighting Approaches for the Calculation of Conformational Variability of Macromolecules from Molecular Simulations

“…Furthermore, the extent and the shape of this landscape may be altered in pathological states [1,2] . This motivates the significant effort devoted to the development of methods to combine experimental and theoretical information for determining the conformational heterogeneity of biomolecules [3–12] . These strategies are either based on restraining, meaning that the experimental data are used directly as a bias to improve the coverage of the conformational space, or on reweighting, meaning that the experimental data are used as a posterior to change the populations of the unbiased ensemble with the aim of improving the agreement between the experimental and the back‐calculated observables [10,13] .…”

“…Earlier in the development of ensemble averaging approaches, MP methods developed first, because smaller ensembles, and thus less computational power were required. Lately, the development of dedicated hardware [30] and enhanced sampling techniques, [31–34] , and the development of improved forcefields, [35] has led to a dramatic improvement of the conformational distributions in terms of both extent and relative populations, shifting the interest towards methods that use a significant share of the prior distribution, like ME methods [12,13,36–41] . To the best of our knowledge, no systematic comparison of different methods using the same datasets has been presented in the literature.…”