< Back to Article

Exploration of the Dynamic Properties of Protein Complexes Predicted from Spatially Constrained Protein-Protein Interaction Networks

Figure 2

Overview of the MCMC PCA algorithm.

A: Initialization of the algorithm involves three steps. The proteins are modeled as spheres (1) and randomly placed on a grid (2). This random placement constitutes as the initial conformation sampled. A score is calculated for this initial complex (3). B: At the beginning of each iteration, three protein is chosen at random (with repetition) to be translated by up to 4 nm in a random direction in x, y and z. The arrow indicates a move chosen by the algorithm. C: We accept the new complex according to the Metropolis-Hasting sampling method.

Figure 2