A Model-Theoretic Approach to Modular Implementation, with Application to Biological Systems

TL;DR

This paper introduces a hierarchical, model-theoretic framework for specifying and analyzing modular systems, especially biological ones, focusing on communication pathways and their effects on system behavior.

Contribution

It presents a formal, hierarchical approach to model the behavior of modular systems with signaling pathways, enabling formal verification and diagnosis of system properties.

Findings

Framework applicable to biological systems with hierarchical modularity

Allows formal proof of system properties and diagnosis

Potential for application in biological system analysis

Abstract

We present a method for logical specification of the behavior of a system based on the specification of its components and for representing and specifying the effect of a class of communication "signaling pathways" among them. Variations in the actions and effects (semantics) of the pathways can be seen in the resulting system. We show how to "enrich" the behavior of the original components by exactly what is provided by the signaling pathways. This has the potential to allow a formal proof of a property of the system as a whole, or a formal diagnosis why a system of communicating modules may not satisfy a desired specification, and if and how it can be "fixed." There are five distinct stages in our view: 1. A description of the connections between potential components at the level of which potential modules are connected. 2. Specific choices for those modules 3. An "elaboration" of the connections among different modules to connections among members of those different modules 4. A description of the semantics of those connections (what they actually do) 5. A "padding" of the modules chosen in step 2 to allow for the action of the connections (step 3), as described in step 4. The formalism is completely hierarchical -- in other words, the formalism works equally for the system as a whole, for each of the modules, and for any larger system in which the original system is a component. As it turns out, for whatever reason -- perhaps evolutionary advantage optimization at the right level of abstraction -- many biological systems are hierarchically modular, with the modules communicating via various pathways. Thus, a formalism such as the one presented here could have application to questions in biology.