Multiscale Computational Modeling in Vascular Biology: From Molecular Mechanisms to Tissue-Level Structure and Function - Multiscale Computer Modeling in Biomechanics and Biomedical Engineering - page 222

importantly, however, CMMs cannot exploit directly the growing knowledge-base
of molecular vascular biology.
3.4.2 ABM Limitations
Agent-based models can explicitly include complex cell-cell interactions, but are
limited when it comes to modeling multiscale phenomena. For example, current
ABMs do not satisfy some of the important constraints on vascular behavior that
stem from classical physics (e.g., conservation of linear momentum). Simulating
whole tissues would also require an extremely large number of individual agents,
which quickly becomes prohibitive computationally, particularly for complex rule-
sets that may include stochastic rules. While stochastic rules can produce a pop-
ulation of results that more closely resembles actual experimental data, including
stochasticity necessitates an even larger number of model runs to converge on an
average output. Finally, single-scale, cell-level ABMs treat the agent as a black
box; intracellular interactions are often included implicitly when rules are derived
from cell-level experiments. When intracellular processes are simple, this is less of
a problem, but as intracellular interactions grow more complex, experimental
approaches to rule development can become intractable. Of course, as with any
model, the ABM rule-set is only as good as the data from which it is derived. Often
the kind of data needed to develop rules for cellular behaviors with complex
multifactorial inputs are not available in the literature.
3.4.3 ISM Limitations
Models of intracellular signaling networks are useful tools in predicting the effects
of signal transduction in a cell in response to a stimulus. These predictions are
made at the cellular level, and may require coupling with higher level models to
infer tissue or organ level function. For example, Ca 2+ signaling in vascular
smooth muscle cells is studied to make predictions about vasomotion in blood
vessels. Vasomotion requires the synchronization of oscillations in the concen-
tration of Ca 2+ in a large group of vascular smooth muscle cells, and gap junctions
are believed to play an important role in this process [ 18 ]. Hence, elements of cell-
cell interactions often need to be coupled to an ISM to convincingly bridge the gap
between single cell and tissue level predictions. Lastly, as with the other types of
mathematical models, determination of the scope and complexity of an ISM during
the developmental stages is not a trivial task. In some cases, the paucity and
confidence of relevant experimental data to be used as model parameters can
directly limit the scope and usefulness of an ISM.