July 7, 2025 | Randles Lab publishes and presents two papers at the International Conference on Computational Science in Singapore this week.
We’re excited to share that members of our lab will be presenting two cutting-edge contributions this week at the International Conference on Computational Science (ICCS 2025). Both papers push the boundaries of what’s possible in multiscale simulation and single-cell analysis, and we look forward to discussing them in person with colleagues from around the world.
Microfluidic Digital Twin for Enhanced Single-Cell Analysis
Samreen T. Mahmud, Wentao Ma, Taylor Thomsen, Chang Chen, Rachel Rex, Andre Lai, Lydia L. Sohn & Amanda Randles
A collaboration with UC Berkeley, this work presents a first-of-its-kind digital twin for the mechano-node-pore sensing (mechano-NPS) platform. By building a virtual replica that unites high-fidelity fluid-dynamics models with biological cell-behavior simulations, we create a powerful tool for device optimization and hypothesis testing. After verifying our fluid-flow simulations against analytical benchmarks, we tuned cellular models to reproduce key experimental observables—most notably cell velocity and the whole-cell deformation index (wCDI). While systematic biases in velocity surfaced, the strong agreement in wCDI confirms the twin’s predictive power. This framework not only enhances mechano-NPS design but also exemplifies how digital twins can revolutionize experimental biology.
Adaptive Physics Refinement for Anatomic Adhesive Dynamics Simulations
Aristotle Martin, Willliam Ladd, Wendy Wu, and Amanda Randles
In “Adaptive Physics Refinement for Anatomic Adhesive Dynamics Simulations,” we tackle the long-standing challenge of explicitly simulating circulating tumor cell (CTC) adhesion across multiple anatomical scales at submicron resolution. Our hybrid CPU–GPU adaptive physics refinement (APR) method dynamically couples a finely resolved adhesive interaction window—capable of traversing complex vessel walls—with a larger fluid domain. Key algorithmic advances include efficient data-movement strategies, a minimized memory footprint, and GPU-accelerated adhesive computations. By applying these techniques to a large microfluidic device, we demonstrate adhesive transport simulations at only a fraction of the cost of fully explicit models, opening the door to new, more realistic studies of metastatic cell behavior.
Join Us at ICCS 2025
If you’re attending, please stop by our sessions to learn more and share your insights. For full details on times, locations, and live streaming options, visit the conference website:
https://www.iccs-meeting.org/iccs2025