Using argument notation to engineer biological simulations with increased confidence

Abstract

The application of computational and mathematical modelling to explore the mechanics of biological systems is becoming prevalent. To significantly impact biological research, notably in developing novel therapeutics, it is critical that the model adequately represents the captured system. Confidence in adopting in silico approaches can be improved by applying a structured argumentation approach, alongside model development and results analysis. We propose an approach based on argumentation from safety-critical systems engineering, where a system is subjected to a stringent analysis of compliance against identified criteria. We show its use in examining the biological information upon which a model is based, identifying model strengths, highlighting areas requiring additional biological experimentation and providing documentation to support model publication. We demonstrate our use of structured argumentation in the development of a model of lymphoid tissue formation, specifically Peyer's Patches. The argumentation structure is captured using Artoo (www.york.ac.uk/ycil/software/artoo), our Web-based tool for constructing fitness-for-purpose arguments, using a notation based on the safety-critical goal structuring notation. We show how argumentation helps in making the design and structured analysis of a model transparent, capturing the reasoning behind the inclusion or exclusion of each biological feature and recording assumptions, as well as pointing to evidence supporting model-derived conclusions.

Keywords: Artoo; argumentation; computational modelling; immune system modelling; simulation.

Figure 1.

(a) A screenshot of the Artoo argumentation tool. This runs in a Web browser window in an up-to-date version of either Chrome or Firefox browser (other browsers do not fully support the required technologies). The File menu provides three options: open a previously developed Artoo argument structure; save an argument structure, export the current argument structure as a PNG image. The Build menu provides options to create, edit and delete nodes (representing argument components); and to create or delete connections among nodes. The View menu operations enable zooming in, zooming out and centring of the argument structure. Nodes in the argument structure can be individually moved using the computer mouse by left clicking and dragging the node. Left mouse clicking and dragging on areas outside of the nodes will drag the entire structure. Right mouse clicking on an argument node (shown in the screenshot on the ‘Assumptions' node) allows access to a node-specific menu to edit or delete the node, add or delete a connection to that node, or collapse and hide the nodes below this node in the tree. A collapsed section of an argument is denoted by a black diamond symbol. (b) Definitions of each node type available in goal structuring notation.

Figure 2.

Top-level of argumentation structure used during the development of a computational tool that captures Peyer's Patch development, as output by our Artoo. The tool enables the developer to capture their claims and evidence using GSN. A claim is made that the simulation is an adequate representation of the biology, and arguments stated that support this claim. In turn, this claim is split into four subclaims. A black diamond notes that the claim has been developed yet is shown in the following figures due to limitations on space. Claim 1.1.3 has been developed in this figure, noting the evidence that simulated cell behaviour at the 12 h time-point is statistically similar to that observed ex vivo. Where a goal is stated, the tool enables the developer to link the goal to the evidence that supports the claim. Here, the figure is demonstrating that this evidence is within selected publications.

Figure 3.

Expanded argumentation structure for claim 1.1.1 in figure 2. The argument was created during the development of a lymphoid tissue development simulator that captures Peyer's Patch development. The claim argues that the biological data against which the simulation is judged are adequate. Where a claim is stated, the tool allows the developer to link the claim to evidence that supports the claim. In the majority of cases, evidence is provided to substantiate the claim. However, where data are unavailable, the claim cannot be substantiated, shown by a blank diamond.

Figure 4.

Expansion of claim 1.1.2 in figure 1. This claim examines the abstractions that have been made and whether these are fit for purpose. In this case, the implementation of chemokines, adhesion factors, the intestine environment and cell signalling is explored.

Figure 5.

Expansion of claim 1.1.4 in figure 1. This claim states that the simulator appropriately replicates the observed emergent behaviour: cell aggregation indicative of Peyer's Patch formation.