Semantics-Based Composition of Integrated Cardiomyocyte Models Motivated by Real-World Use Cases.

Maxwell L Neal,Kenneth Tran,John H Gennari,Karam G Kim,Christopher T Thompson,Daniel L Cook,Brian E Carlson,Edmund J Crampin,Ryan C James,Neil R Smalheiser

doi:10.1371/journal.pone.0145621

Abstract

Semantics-based model composition is an approach for generating complex biosimulation models from existing components that relies on capturing the biological meaning of model elements in a machine-readable fashion. This approach allows the user to work at the biological rather than computational level of abstraction and helps minimize the amount of manual effort required for model composition. To support this compositional approach, we have developed the SemGen software, and here report on SemGen’s semantics-based merging capabilities using real-world modeling use cases. We successfully reproduced a large, manually-encoded, multi-model merge: the “Pandit-Hinch-Niederer” (PHN) cardiomyocyte excitation-contraction model, previously developed using CellML. We describe our approach for annotating the three component models used in the PHN composition and for merging them at the biological level of abstraction within SemGen. We demonstrate that we were able to reproduce the original PHN model results in a semi-automated, semantics-based fashion and also rapidly generate a second, novel cardiomyocyte model composed using an alternative, independently-developed tension generation component. We discuss the time-saving features of our compositional approach in the context of these merging exercises, the limitations we encountered, and potential solutions for enhancing the approach.

Highlights

While biomedical researchers have the computational power to express complex hypotheses about biological systems using quantitative biosimulation models, repurposing previouslypublished models to compose larger, multiscale systems remains a largely manual, time-consuming and error-prone task
To validate our version of the PHN model, we compared our simulation results for intracellular calcium, membrane voltage, and active myocyte tension to those from the model published by Terkildsen et al We investigated the mathematical differences between the versions that accounted for any deviations between the results
These calcium dynamics are controlled by the Hinch calcium mass balance equations, and they in turn produce active myocyte tension development according to the Niederer model formulation

Summary

Introduction

While biomedical researchers have the computational power to express complex hypotheses about biological systems using quantitative biosimulation models, repurposing previouslypublished models to compose larger, multiscale systems remains a largely manual, time-consuming and error-prone task. Developing such tools is challenging because, for example, modelers do not all encode their models in the same language and they do not always employ the same system of physical units Another critical challenge is that systems-level hypotheses differ between research groups; two modelers may conceptualize the same system at different levels of granularity, with different sets of biological role players, and with different input/output arrangements between interacting components. There is a need for a flexible model composition approach that accommodates the various ways researchers conceptualize a biological system, and one that scales across modeling languages [1] To meet these needs we have developed the SemSim (semantic simulation) model description format [2,3], which describes a biosimulation model in terms of its biological meaning formally linked to the model’s specific mathematical implementation. We currently encode SemSim models in OWL [4], but other knowledge representation formats might be used

Objectives

Methods

Results

Conclusion