Comparing state-of-the-art approaches to back-calculate SAXS spectra from atomistic molecular dynamics simulations

Small-angle X-ray scattering (SAXS) experiments are arising as an effective instrument in the structural characterization of biomolecules in solution. However, they suffer from limited resolution, and complementing them with molecular dynamics (MD) simulations can be a successful strategy to obtain information at a finer scale. To this end, tools that allow computing SAXS spectra from MD-sampled structures have been designed over the years, mainly differing in how the solvent contribution is accounted for. In this context, RNA molecules represent a particularly challenging case, as they can have a remarkable effect on the surrounding solvent. Herein, we provide a comparison of SAXS spectra computed through different available software packages for a prototypical RNA system. RNA conformational dynamics is intentionally neglected so as to focus on solvent effects. The results highlight that solvent effects are important also at relatively low scattering vector, suggesting that approaches explicitly modeling solvent contribution are advisable when comparing with experimental data, while more efficient implicit-solvent methods can be a better choice as reaction coordinates to improve MD sampling on-the-fly. Graphic abstract