Skip to main content
. Author manuscript; available in PMC: 2024 May 1.
Published in final edited form as: J Mol Biol. 2023 Feb 16;435(9):168014. doi: 10.1016/j.jmb.2023.168014

Fig.1. Building Gaussian models from neutral state density maps.

Fig.1.

(A) Resource usage comparison between the voxel based representation and the GMM based one. Note the y axis in the plot is in log scale. (B) Workflow for building a neutral state GMM from the corresponding density map. The first round of training matches the output GMM from the decoder to the coordinates of the molecular model, and the second round of training maximizes the similarity between the projections of the density map and the projections of the GMM.