Summary: State-of-the-art global optimization (GO) algorithms are applied to computation of 3-D virus reconstructions from solution x-ray scattering data by minimizing a weighted least squares error. Starting with random uniformly-distributed initial conditions, high quality reconstructions are achieved using GO algorithms. Similar quality reconstructions using classical multistart require initial conditions that incorporate more detailed a priori knowledge which is not always available. The behavior of the GO algorithms with respect to the parameterization of the error function and with respect to the accuracy of the quadrature rule embedded in the calculation of the regressor are described. Several future directions in terms of structural biology and GO algorithms are discussed.
For the entire collection see [Zbl 1103.90011].