Principal component analysis

The identification of relevant motions in MD trajectories is carried out using the PCA that is able to filter global, collective, and slow motions (relevant) from local and fast motions (not relevant). This operation is accomplished by projecting the trajectories onto the eigenvectors to give the principal components. Eigenvectors are the columns of the diagonalized mass‐weighted covariance matrix of the atomic positional fluctuations (Amadei et al, 1993), which is calculated on protein backbone atoms. The first few principal components represent the most relevant motions and the first one contains the largest mean‐square fluctuation.