Skip to main content
Fig. 2 | BMC Bioinformatics

Fig. 2

From: Complementing sequence-derived features with structural information extracted from fragment libraries for protein structure prediction

Fig. 2

The overall pipeline of building wGMM models and protein-specific potentials from fragment libraries. a The overall pipeline that transforms fragment libraries into potentials for gradient descent-based protein folding. Fragments of variable lengths are first cut into a series of 7-residue fragments by smoothing. The color of a fragment denotes its “source” fragment (i.e., three fragments with 7 residues colored in blue in the second subfigure are all cut from the 9-residue fragment colored in blue in the first subfigure). Protein properties are extracted from the smoothed fragment library and fitted by weighted Gaussian mixture models (wGMM models). wGMM models are converted into potentials and utilized in SAMF [22] for protein folding. be Visualization of the wGMM models for ϕ (b), ψ (c), θ (d), τ (e) of the first residue of 67th position of T0969-D1. The red lines mark the corresponding properties of the native protein structure

Back to article page