In Pursuit of Total Reproducibility

Abstract

The vast majority of scientific contributions in the field of computational systems biology are based on mathematical models. These models can be broadly classified as either dynamic (kinetic) models or steady-state (constraint-based) models. They are often described in specific markup languages whose purpose is to aid in the distribution and standardization of models. Despite numerous established standards in the field, reproducibility remains problematic due to the substantial effort required for compliance, diversity of implementations, and the lack of proportionate rewards for researchers. This article explores the application of event sourcing - a software engineering technique where system state is derived from sequential recorded events - to address reproducibility challenges in computational systems biology. Event sourcing, exemplified by systems like git, offers a promising solution by maintaining complete, immutable records of all changes to a model. Through examples including leader and follower applications, local and remote computation, and contribution tracking, this work demonstrates how event-sourced systems can automate standards compliance, provide comprehensive audit trails, enable perfect replication of processes, facilitate collaboration, and generate multiple specialized read models from a single event log. An implementation of the outlined principles has the potential to transform computational systems biology by providing unprecedented transparency, reproducibility, and collaborative capabilities, ultimately accelerating research through more effective model reuse and integration. An event-sourced approach to modeling in computational systems biology may act as an example to related disciplines and contribute to ending the reproducibility crisis plaguing multiple major fields of science.