Gaussian mixture model for coarse-grained modeling from XFEL

• DOI: 10.1364/oe.26.026734
• Publication Date: 2018-09-28T18:11:24Z
• Authors: Tetsuro Nagai; Yuki Mochizuki; Yasumasa Joti; Florence Tama; Osamu Miyashita
• Keywords: 0303 health sciences; 03 medical and health sciences

We explore the advantage of Gaussian mixture model (GMM) for interpretation single particle diffraction patterns from X-ray free electron laser (XFEL) experiments. GMM approximates a biomolecular shape by superposition distributions. As Fourier transformation can be quickly performed, we efficiently simulate XFEL approximated structure models. report that resolution accurately reproduce is proportional to cubic root number Gaussians used in modeling. This behavior attributed correspondence between adjustable parameters and amount sampling points space. Furthermore, GMMs successfully perform angular assignment detect conformational variation. These results demonstrate serve as useful coarse-grained models hybrid approach