Extended Data Fig. 2: Highlight of the main differences found in the consensus landscape estimated for the EMPIAR 10028 dataset.
From: Merging conformational landscapes in a single consensus space with FlexConsensus algorithm
The top panel shows the regions where Zernike3D and HetSIREN disagree due to the presence of a strong compositional change that Zernike3D cannot capture. The bottom panel shows a strong difference in the periphery of the landscape, mainly associated with wrongly estimated conformations with increasing noise in the case of HetSIREN. Since Zernike3D cannot add noise to the conformation it estimates, its periphery is closer to the central region of the consensus space, as shown in the Figure.
