Establishing Digital Twins for High Throughput Cellular Analysis in Whole Blood

PI Amanda Randles, Duke University
Co-PI Lydia Sohn, University of California Berkeley
Randles INCITE Graphic

Digital twin of a mechano-NPS microfluidic device. Image Courtesy of the Randles Lab at Duke University

Project Summary

This INCITE project seeks to address the major challenges facing cellular simulations to allow cancer researchers to quickly identify potentially deleterious mechano-phenotypes. 

Project Description

It has been well established that mechanical properties of cells change with differentiation, chronological age, and disease state. Specifically, a positive correlation between increased cell deformability and heightened metastatic potential has been observed for breast cancer cells. More broadly, the changes caused by many cancers within the cell are manifest in measurable biophysical properties, such as optical deformability and size. Being able to screen and accurately characterize many cells would allow cancer researchers to quickly identify potentially deleterious mechano-phenotypes. 

To better understand this concept, this team developed a series of in vitro based techniques. Atomic-force microscopy (AFM) and micropipette aspiration are the gold standard for performing mechanical measurements on cells. However, they are limited in terms of throughput. Given these drawbacks, these researchers have turned to microfluidic platforms to extract mechanical properties from specific cell types with high throughput. Mechano- phenotyping platforms can now quantify four fundamental biophysical properties: diameter, resistance to compressive deformation, transverse deformation, and recovery from deformation in a label-free manner. 

The team’s work addresses three key challenges facing cellular simulations. First, being able to establish a precise and validated method for a digital twin of the microfluidic device. Second, the team will extend this capability by creating a framework for robustly capturing cellular behavior across the ensemble of potential red blood cell configurations. Finally, the team will set a computationally optimized method by integrating this framework with their adaptive physics refinement method that enables cellular resolution to be captured over large domains. Together, these advances will allow a robust digital twin for a wide range of microvascular and microfluidic applications for systems targeting whole blood analysis