Elucidation of molecular kinetic schemes from macroscopic traces using system identification

• Published in: PLoS computational biology 2017-02, Vol.13 (2), p.e1005376-e1005376
• Main Authors: Fribourg, Miguel, Logothetis, Diomedes E, González-Maeso, Javier, Sealfon, Stuart C, Galocha-Iragüen, Belén, Las-Heras Andrés, Fernando, Brezina, Vladimir
• Format: Article
• Language: English
• Subjects: Algorithms; Biology and Life Sciences; Biophysics; Biopolymers - chemistry; Biopolymers - metabolism; Cell interactions; Cell Physiological Phenomena; Computer Simulation; Enzyme kinetics; Funding; Kinetics; Medicine; Medicine and Health Sciences; Metabolic Clearance Rate; Models, Biological; Models, Chemical; Observations; Physical Sciences; Physiology; Research and Analysis Methods; Signal Transduction - physiology
• ISSN: 1553-7358; 1553-734X; 1553-7358
• DOI: 10.1371/journal.pcbi.1005376

Overall cellular responses to biologically-relevant stimuli are mediated by networks of simpler lower-level processes. Although information about some of these processes can now be obtained by visualizing and recording events at the molecular level, this is still possible only in especially favorable cases. Therefore the development of methods to extract the dynamics and relationships between the different lower-level (microscopic) processes from the overall (macroscopic) response remains a crucial challenge in the understanding of many aspects of physiology. Here we have devised a hybrid computational-analytical method to accomplish this task, the SYStems-based MOLecular kinetic scheme Extractor (SYSMOLE). SYSMOLE utilizes system-identification input-output analysis to obtain a transfer function between the stimulus and the overall cellular response in the Laplace-transformed domain. It then derives a Markov-chain state molecular kinetic scheme uniquely associated with the transfer function by means of a classification procedure and an analytical step that imposes general biological constraints. We first tested SYSMOLE with synthetic data and evaluated its performance in terms of its rate of convergence to the correct molecular kinetic scheme and its robustness to noise. We then examined its performance on real experimental traces by analyzing macroscopic calcium-current traces elicited by membrane depolarization. SYSMOLE derived the correct, previously known molecular kinetic scheme describing the activation and inactivation of the underlying calcium channels and correctly identified the accepted mechanism of action of nifedipine, a calcium-channel blocker clinically used in patients with cardiovascular disease. Finally, we applied SYSMOLE to study the pharmacology of a new class of glutamate antipsychotic drugs and their crosstalk mechanism through a heteromeric complex of G protein-coupled receptors. Our results indicate that our methodology can be successfully applied to accurately derive molecular kinetic schemes from experimental macroscopic traces, and we anticipate that it may be useful in the study of a wide variety of biological systems.