Abstract
Advanced simulations within biophysical applications ask for advanced algorithms and implementations which are running efficiently on massively parallel high performance computers. The software framework UG fulfills these preconditions. Therefore, we present insight into the experimental basics, the modelling and simulation details, and the biophysical meaning of the estimation of the diffusion constant of a major player in the replication of the genetic information of the Hepatitis C virus (HCV), namely the NS5A viral protein. NS5A movement is restricted to the surface of the Endoplasmatic Reticulum (ER, a medusa-hair like important cell compartment). Hence, the dynamics of NS5A are described by surface PDEs (sPDE) which mimic experimental FRAP (fluorescence recovery after photobleaching) time series data. The sPDE computations were performed with UG upon large unstructured grids representing realistic reconstructed ER surfaces. We explain the context of the parameter estimations which asked for a substantial amount of single sPDE evaluations which we performed on the HLRS Stuttgart Hermit and Hornet supercomputers for various experimental data sets and for various geometric setups. This enabled us to derive valid final values for the diffusion constant of NS5A on the ER surface. The estimated diffusion constant values are intended to enter spatio-temporal resolved models of HCV replication dynamics at a cellular level.
Original language | English (US) |
---|---|
Title of host publication | High Performance Computing in Science and Engineering '15 |
Subtitle of host publication | Transactions of the High Performance Computing Center, Stuttgart (HLRS) 2015 |
Publisher | Springer International Publishing |
Pages | 641-657 |
Number of pages | 17 |
ISBN (Electronic) | 9783319246338 |
ISBN (Print) | 9783319246314 |
DOIs | |
State | Published - Jan 1 2016 |
Externally published | Yes |
ASJC Scopus subject areas
- General Computer Science
- General Physics and Astronomy
- General Mathematics
- General Chemistry