An angle based variant of Dynamic cross correlation metric

Molecular dynamic simulation is a powerful tool for finding dependencies in biological molecules. Correlation matrices can map out the nature of correlated movements performed by different regions within the molecule during their solution state simulations and provide a critical tool in understanding the interactions between different residues in a structure. Here we present a methodology to extend a popular method for finding correlation by addressing to known problems with it.Dynamic cross correlation metric (DCCM) also known as Pearson’s coefficient [2] is the most commonly used method for finding correlation in data obtained from Molecular dynamics(MD) simulations. DCCM has well documented fallacies, which are fundamentally related to its closed definition of correlated motion, which focuses on linear correlation of motion. A description of these are found in [1], [4].