Biomedical Engineering Reference
In-Depth Information
increased load on the heart causes left ventricle hypertrophy and left ventricle
failure. Alternatively, we can theorize about ''top-down'' effects. For example,
how does mechanical shear force experienced by the ECs impact NO signaling and
PDGF expression, which in turn affect SMCs and overall wall mechanics? Cre-
ating models to predict these types of outcomes will save time, money, and
potentially lives.
Of course, the aspiration to unite ISM-ABM-CMM is met with considerable
challenges, not the least of which is the requirement for additional computing
power. Assuming the technical challenges can be overcome by advances in
computing (e.g., parallelization, grid computing, and cloud computing), one must
address the conceptual challenges in multiscale model design. The final section of
this chapter will delve deeper into these challenges and suggest opportunities for
innovation in multiscale modeling.
5.2 Challenges
As noted by the 1998 Bioengineering Consortium (BECON) Report of the U.S.
National Institutes of Health,
The success of reductionist and molecular approaches in modern medical science has led
to an explosion of information, but progress in integrating information has lagged …
Mathematical models provide a rational approach for integrating this ocean of data, as
well as providing deep insight into biological processes.
Whereas the need remains to develop more robust and faithful models at all
scales (macro, micro, nano), we submit that there is a pressing need to develop
approaches that integrate such models across diverse scales. Indeed, anticipating
the challenges of multiscale modeling should influence the development of models
at each scale for they will need to interface with the other models. Toward this end,
we suggest here the following particular challenges that deserve our immediate
attention.
There are several computational languages used to run numerical analysis (e.g.,
Matlab, Maple, Mathematica, Java Virtual Machine, FORTRAN, and C++,). Thus,
a logistical challenge may arise when models, at different scales, are programmed
with different languages. We proposed herein using text files as inputs/output
because all our modeling platforms can read and write text files, however this
process is time consuming and cumbersome. Therefore finding patches or proper
interfaces between multiple numerical analysis software remains a challenge. In
addition, iterative simulations may take days to complete and require considerable
memory on a personal computer. Consequently large-capacity databases and fast
processors/parallel systems may be required to render the computational process
tractable. After a multiscale program is completed, finding ways to distill and
partition model findings into digestible chunks that is are easy to disseminate and
publish may be a challenge.
Search WWH ::
Custom Search