X-ray crystallography is the most powerful method to obtain the structural of biological molecules if the “phase problem” can be solved for the molecules under study. The phase problem arises from the loss of phase information in diffraction experiment. In all the solutions of the phase problem, the direct method is the only one that does not require additional experimental data or knowledge of homologous structures. It can determine the phase information directly from the observed structure factor magnitudes or intensities. However, the direct phasing method has limitations when applying to macromolecule. It is only applicable in molecules with up to about 1000 non-H atoms and requires ultra-high resolution (the Sheldrick's 1.2 Å rule) diffraction data that is not available in most protein crystallography experiments. To overcome the two limitations, here we propose a coarse-grained direct phasing method. This thesis will focus on how to break the 1.2 Å resolution requirement.