Main Article Content
In this paper, we studied the efficiency of defining the subspace of the slow protein motion for describing long time scale dynamics. To address the problem of the correlations between the slow and the fast modes, we investigated the separation of time scales provided by the Principal Component Analysis (PCA) method. In our study, two systems have been discussed, the C2 fragment of the G protein (56 amino acids) and the Fc domain of IgG protein (206 amino acids). Our results indicated that only little convergence is obtained at a sub nanosecond time scale for large proteins, whereas pronounced convergence sets in for small proteins at the 5 ns time scale. This finding, together with the demonstrated separation of time scales, suggests that PCA provides indeed suitable subspaces for a dimension reduced description of protein dynamics in long time scales.