Multiscale Computational Modeling in Vascular Biology: From Molecular Mechanisms to Tissue-Level Structure and Function - Multiscale Computer Modeling in Biomechanics and Biomedical Engineering - page 227

refinement. This process allows one to refine the objectively bounded model
parameters in an attempt to minimize the error between common metrics between
both models for simple situations. This step is imperative to synchronize both
models across scales and to yield predictions that are closer to in vivo observa-
tions. This is accomplished not by compromising the strengths of either model, but
by tuning each model so that it is influenced by the strengths of the other. For
example, the ABM and CMM both have the ability to predict the amounts of
collagen and smooth muscle under simple situations; therefore by instituting a
genetic algorithm, or any preferred error minimization technique, the difference
between these common metrics, at each time point, can be minimized by varying
the set of parameters after each full simulation.
5 Future Directions
5.1 Potential Mulitscale Model
A compelling goal in developing multiscale models is to extend the models to both
higher and lower levels of scale, with the motivation being that with every added
level of scale, one gains even greater flexibility with regards to hypothesis testing,
achieving biological relevance, and incorporation of disparate data sets. Starting
with the multiscale CMM-ABM, a natural extension is to conjoin an intracellular
signaling model (ISM) that can simulate events on a much shorter timescale and
account for phenomena within the cell that ultimately impact cell behaviors and
outcomes at the tissue level. We have begun to conceptualize a three-tiered
multiscale model (ISM-ABM-CMM) of vascular growth and remodeling in the
arterial wall (Fig. 5 ), and this section will briefly summarize some of the enabling
components.
The relevant biology of vascular growth and remodeling during human disease
occurs over years; however, biological phenomena contributing to disease pro-
gression over this timeframe occur across incremental time steps that range from
seconds to weeks, depending on the spatial scale. Specifically, signaling events in
the ISM have a timeframe on the order of seconds to minutes, multi-cell phe-
nomena in the ABM have a timeframe of minutes to hours, and tissue-level
continuum phenomena in the CMM have a timeframe of hours to days or months.
Therefore, in the construction of the multiscale model we envision a temporal
decomposition strategy to couple across scales (Box 1).