Although CICR is a regenerative process, calcium from the SR is not depleted during CICR. So, there is a calcium-dependent inactivation, in addition to calcium-activation. Another property is the gradedness of calcium elevations, i.e. not in an all-or-none manner. This raises another paradox, and was explained by the local control theory, i.e. CICR occurs at "local" domains and the global calcium transient is the recruitment of all local elevations [Cheng&Lederer, Bers]. The local elevations occur at microdomain areas known as calcium release units (CRUs). A proper mechanistic explanation of this calcium dynamics require taking into account the realistic number of CRUs in a single ventricular myocyte, i.e. 20000 CRUs. Another challenge is that this local elevation occurs very fast in time (~25-30ms) and confine to a tiny space (2um in diameter). The investigation of spatial distribution of CRUs may lead to insights to calcium dynamics, which can help answering many intriguing scientific questions, e.g. how T-tubule disruption affect to normal functions of the heart, what is the mechanism of calcium-entrained arrhythmias.
With that purpose in mind, at Jafri Lab, I'm conducting research in developing
3D spatiotemporal rat ventricular myocyte model.
The ultimate purpose would be using a system biology approach to
build a mechanistic whole-cell model.
So far, I've developed a highly efficient
Ultra-fast Markov-chain Monte-carlo method (patent-pending), which allows rapid simulation of whole-cell model containing 20,000 CRUs.
This novel method provides a powerful tool for performing stochastic simulation.
Under the tutelage of my advisor, I'm also interested in
high performance computing, especially the emerging
General-Purpose GPU (GPGPU) technology from nVidia - Compute Unified Device Architecture (CUDA).
(CUDA) Fortran, combining with (CUDA) C/C++, is being used to develop our model.