This Title All WIREs
How to cite this WIREs title:
WIREs Comp Stat

Agent‐based modeling and biomedical ontologies: a roadmap

Full article on Wiley Online Library:   HTML PDF

Can't access this content? Tell your librarian.

Abstract The translational dilemma represents a foundational challenge for the biomedical research community. Addressing the dilemma will require an enhancement in the throughput capacity of the expression and evaluation of mechanistic hypotheses. Doing so will require the ability to place biomedical knowledge into a format where hypotheses can be readily instantiated such that the dynamics inherent to biological systems can be represented, and also technological enhancement of the generation of such dynamic models. We suggest that the former goal can be approached by using the meta‐structure of agent‐based models (ABMs) to integrate different knowledge hierarchies currently represented with bio‐ontologies and increase the expressiveness of formal knowledge representation to account for mechanistic biological rules. The development of an agent‐based modeling format (ABMF) will provide a bridge between ontological knowledge representation and methods for modeling and simulation (M&S). We further suggest that the latter goal of process enhancement can be targeted by the development of intelligent computational agents that can concatenate bio‐ontologies with M&S ontologies to semiautomate the generation of bio‐simulations. We believe that the application of these two complementary approaches would address the foundational nature of the current throughput bottleneck in the scientific cycle. WIREs Comp Stat 2011 3 343–356 DOI: 10.1002/wics.167 This article is categorized under: Statistical Models > Agent-Based Models Applications of Computational Statistics > Computational and Molecular Biology Software for Computational Statistics > Artificial Intelligence and Expert Systems Statistical Models > Simulation Models

The scientific cycle. A schematic representation of the classical cycle of observation, interpretation, hypothesis formation, and experiment. Note that the cycle is iterative and can be considered as a work‐flow diagram. (Reprinted with permission from Ref 2. Copyright 2010 AAAS)

[ Normal View | Magnified View ]

Nested agent‐based modeling format (ABMF) behavior hierarchies from an ABM of gut‐lung inflammation. This is a similarly abstracted representation of the behavioral descriptors, or verbs, present in the ABM of gut‐lung inflammation.18 Note that while these behavior descriptions are associated to their components using first order predicate logic, they form the labels for the rules present and expressed as part of the ABM structure. Also note that the need to have two separate diagrams (Figure 5 and 6) to represent the ABMF structure points to orthogonal nature of the description hierarchies contained in the ABMF.

[ Normal View | Magnified View ]

Nested agent‐based modeling format (ABMF) attribute hierarchies from an agent‐based model (ABM) of gut‐lung inflammation. This is an abstracted representation of an existing multihierarchical ABM of the gut‐lung axis of acute inflammation18 that focuses on the attributes, or noun/adjectives in the ABM description. Note that there are two orthogonal hierarchies of organization: the vertical axis that denotes the trans‐hierarchical representation capacity of an ABM (Figure 3) whereas the horizontal axis demonstrates a parent‐child inheritance hierarchy associated with the object‐oriented paradigm and ontologies.

[ Normal View | Magnified View ]

A schematic description of an agent‐based modeling format (ABMF) module. The ABMF module incorporates three hierarchical levels of representation centered on the simulation agent level. This level corresponds to the classical Agent‐level in an ABM, where the system‐level corresponds to agent population behavior (including so called ‘emergent’ phenomenon), whereas the lowest level of organization, generative mechanisms, corresponds to agent‐rules. It should be noted that agent‐rules can be any formal model system, including another ABM. This property gives the ABMF a potentially recursive structure that allows nesting of ABMF modules. Also, as the generative rules can be any type of mathematical model, the ABMF is also capable of being a pathway to hybrid computational models that concurrently employ multiple M&S methods.

[ Normal View | Magnified View ]

The mapping between hierarchies of biological organization, research community structure and an agent‐based model (ABM). This diagram demonstrates a series of hierarchies present in biological systems, the research communities established to study them, and an ABM. Note that the hierarchies are nested in the biological system and the ABM, reflecting the trans‐hierarchical coupling seen in both systems. Alternatively, the research community structure is disparate and compartmentalized, due to both social and epistemological factors. (Reprinted with permission from Ref 24. Copyright 2006)

[ Normal View | Magnified View ]

The current imbalance in the scientific cycle. As a result of the disproportionate growth in the ability to generate, collect, and correlate data, a process bottleneck has developed at the step involving the evaluation of hypotheses via experiment. This bottleneck underlies the translational dilemma by making it impossible to systematically and efficiently establish trust in putative mechanisms such that they can be targeted for therapeutic control. (Reprinted with permission from Ref 2. Copyright 2010 AAAS)

[ Normal View | Magnified View ]

Browse by Topic

Statistical Models > Simulation Models
Software for Computational Statistics > Artificial Intelligence and Expert Systems
Applications of Computational Statistics > Computational and Molecular Biology
Statistical Models > Agent-Based Models

Access to this WIREs title is by subscription only.

Recommend to Your
Librarian Now!

The latest WIREs articles in your inbox

Sign Up for Article Alerts