Project Group: Dynamic Modelling

Margot Riggi

Dynamic modeling of complex molecular mechanism

As the scientific community gathers diverse data describing how life mechanisms occur across multiple spatial and temporal scales, a key challenge for modern biology is to integrate the richness of this information into coherent models that reflect our rapidly evolving state of knowledge.

We use and advance complementary computational approaches to build and share intuitive, dynamic models of complex molecular mechanisms.

1. 3D animation

That polished 3D animations are a powerful means to engage lay audiences in Science is rarely contested. What is still generally underappreciated is how the process of creating an animation can also play an integral role in research: the consolidation of complementary structural and dynamic datasets into intuitive models often reveals gaps in knowledge, while allowing hypothesis exploration and promoting scientific discourse.

We work in collaboration with researchers to create animated models that capture current hypotheses on various molecular mechanisms. We are also interested in developing methods to facilitate the integration of new types of data and to increase the transparency of the resulting models.

The animations below were created as part of the Iwasa Lab at the University of Utah.

In collaboration with Elizabeth Villa (UCSD). Read more: Laughlin & al, Nature, 2022.

In collaboration with Kelly Lee (University of Washington). Read more: Kephart, SM. & al, Trends in Biochemical Sciences, 2024.

In collaboration with Peter Shen (University of Utah). Read more: Wang S. & al, Molecular Cell, 2023.

2. Agent-based simulation

Simulation approaches go beyond integration of already available data and are able to make predictions that can be experimentally tested. However, understanding complex molecular mechanisms requires spanning broad scales in space (nm to um) and even greater in time (ps-ns to ms-sec). Different methods can be used to carry out simulations on specific ranges of these scales, but all require tradeoffs to be made to balance spatial resolution and simulation length with computational cost. By translating outputs from one level of representation into inputs for a neighboring one, multiscale modeling aims to connect across scales and understand how changes at different levels influence each other. However, linking representations to bridge the molecular and cellular scales remains challenging because of computational cost and the relative lack of experimental data at this intermediate scale. This calls for novel, efficient mesoscopic modeling methods.

Agent-based modeling (ABM) is a stochastic and bottom-up approach to simulate a complex system and its emergent properties from the perspective of its individual components, termed “agents”. Instead of being directly driven by biophysical forces, each agent behaves according to a set of rules that can be derived from very diverse information. Some of these rules can still rely on biophysical equations, while others may be defined from outputs of simulations carried out at lower scales, or stem from experimental data. Importantly, this approach achieves extended time scales, which allows for direct comparison to experimental results for both validating the models and testing further predictions.

While several ABM tools already exist, we are developing a new workflow to build intuitive 3D simulations that explicitly incorporate intramolecular conformational changes in addition to Brownian diffusion and molecular interactions.