Petri nets formalism facilitates analysis of complex biomolecular structural data

Molecular dynamics (MD) simulation is a popular method of protein and nucleic acids research. Current MD output trajectories are huge files and therefore they are hard to analyze. Petri nets (PNs) is a mathematical modeling language that allows for concise, graphical representation of complex data. We have developed a few algorithms for PNs generation from such large MD trajectories. One of them, called the One Place One Conformation (OPOC) algorithm, is presented in a greater detail. In the OPOC algorithm one biomolecular conformation corresponds to one PN place and a transition occurring in PN graph is related to a change between biomolecules conformations. As case studies three simulations are analyzed: an enforced steered MD (SMD) dissociation of a transthyretin protein tetramer into dimers, the SMD dissociation of an antibody-antigen complex and a classical MD simulation of transthyretin. We show that PNs reproduce events hidden in MD trajectories and enable observations of the conformational space features hard-to-see by the other clustering methods. Thus, a fundamental process of biomolecular data classification may be optimized using the PN approach.