Page 162 - MODES of EXPLANATION

Basic HTML Version

The researchers were able to do many things once they understood the dynamical
behavior of the system. A purely mathematical understanding allowed them track and control
the propagation of patterns throughout the system. It gave them affordances for action, but
not a mechanistic explanation. There is still no explanation for what is causing the bursting
phenomena, but they were able to make an enormous amount of scientific progress without it.
Understanding without explanation in bio-engineering
The second case, an integrated systems biology lab, is largely populated by engineers with
little biological background; this is true among integrated systems biology labs more
generally. Yet these engineers are very successful at modeling biological practices. Where do
they get their biological understanding and background? They learn the systems through their
model-building processes. The skills and judgments they are using in their modeling
processes are related to the mathematical relations among variables. That is, they try to
understand the systems dynamics. So their understanding of the biological phenomena is at a
level of mathematical understanding, which could not reasonably be called an explanation.
Lab A aims to design vascular tissue replacements for the human circulatory system.
To do this it needs to understand the mechanics of arteries: how does the pressurized flow of
blood influence the design of arteries? Researcher A1 worked on atherosclerosis. She
discovered that in an artery, oscillatory sheer stress produces inflammation whereas laminar
flow does not. The process of this discovery started when some benchtop molecular
biologists gave her a little piece of a pathway to model. Such biologists rarely provide
enough data for the engineers to provide a model, so A1 conducted a very deep literature
search, gathering all the articles she could that seemed to relate, following up references, and
so on. Remember, this is a person who knew nothing about the system she was working on
when she started. One day she might be working on cancer, tomorrow on atherosclerosis;
there was very little continuity, biologically speaking, between her projects. Nevertheless,
what engineers such as her do possess is skill and judgments in building a model. In addition
to data, engineers working on integrated systems biology often rely on physical simulations.
A1 used a
flow channel device
that the lab referred to as the “flow loop.” This allows
researchers to simulate conditions inside a blood vessel:
“[A]s engineers, we try to emulate that environment [inside an artery], but we also try
to eliminate as many extraneous variables as possible. So we can focus on the effect
of one or perhaps two, so that our conclusions can be drawn from the change of only
one variable… [The flow loop provides] a way to impose a very well-defined shear
stress across a very large population of cells such that their aggregate response will be
due to [the shear stress] and we can base our conclusions on the general response of
the entire population.” (Source A10, Osbeck, et al, 2010)
A1’s original model contained five variables. After she had moved back and forth between
simulation and building pathways, her final model contained fourteen variables.
5