Computer modeling of solution X-ray scattering intensity for biomacromolecules
View/ Open
Author
Tong, Dudu
Date of Issue
2016-12-01School
School of Biological Sciences
Abstract
Increasing amounts of scattering data are obtained from high-throughput Solution X-ray scattering (SXS) experiments, including small-angle X-ray scattering (SAXS) and wide-angle X-ray scattering (WAXS). There is a great demand of computational methods that can retrieve useful structural
information or build structure model from these data. We have proposed
substantial improvements to current methods that model the scattering profiles from protein structure at both atomistic scale and coarse-grained (CG) scale. In addition, our coarse-grained approach can be conveniently applied to structure optimization based on target scattering intensity.
Firstly, a fast Fourier transform (FFT) based orientational average method is
proposed to improve the computational efficiency of modeling scattering
profiles using an atomistic protein structure representation, especially in case of considering explicit hydration water molecules. Comparing with the popular
spherical average method, our method will become more efficient for systems
with more than 3000 atoms. Moreover, the computational time of our
FFT-based method remains nearly unchanged as the system size increases,
making it suitable for very large protein complexes.
CG representations are also widely used to improve the computational
efficiency of theoretical scattering intensity computation. Given the importance
of accuracy for CG approaches, we have proposed the electron density
matching (EDM) method to parameterize the CG form factors. Comparing with
the CG form factors used in literature, our EDM-derived ones result in better
agreement to atomistic scattering intensities. Furthermore, the resulting CG
xxform factors are shown to reproduce the experimental scattering profiles well by including the contribution of hydration layer and the correction of protein excluded volume.
Finally, in order to perform structure modeling with our EDM-derived CG form
factors, we have proposed an implicit hydration term to take account the
contribution of the hydration layer scattering. This term is only related to the
surface accessible solvent area (SASA) of protein atoms, making our formulation to evaluate scattering intensity analytically differentiable to the
protein coordinates. The implicit hydration term is fitted to best reproduce the
overall scattering intensity computed using explicit hydration water molecules.
It is shown that the conjugate gradient structure optimization based on the target scattering intensity can produce final molecular structures very close to the known target structure.
Subject
DRNTU::Science
Type
Thesis
Collections