Cellular phenotype modeling of the long QT syndrome gene supported by distributed/parallel computation

Characterizing the cellular phenotype of the mutated long QT syndrome (LQTS) gene is the bridging approach to relating the molecular dysfunction to the electrophysiological abnormalities. In this study, we model the cellular functional characteristics of the delayed rectifier K+ channel, I
kr
, responsible for the LQTS subtype 2. The Levenberg-Marquardt algorithm was used to estimate the transition rate function of the Markov model of the channel kinetics, whereby the calculation of the gradient matrices of the merit function was decomposed and implemented by use of the message passing interface. With up to 32 CPUs, the parallel efficiency of the parallel computation was greater than 89%. The kinetic models of the widetype and mutated wide-type/p.Pro872fs channels were established with a significant improvement in computing efficiency.