Illustration of the diverse and often unpredictable phenomena that cells encounter while circulating through the circulatory system. This includes a range of fluid dynamics, from smooth laminar flow to chaotic turbulent flow, passage through narrow and constricted vessels, interactions with plaques and vessel irregularities, and the dense, crowded environment created by surrounding blood cells, platelets, and other suspended components within the bloodstream.

While mutations are the traditional drivers of evolution, recent decades have highlighted the critical role of physical forces in shaping evolutionary processes. This is most evident during an organism’s developmental stage; to survive, new cells must evolve mechanisms to sense and respond to stimuli such as hydrostatic pressure, shear stress, and the geometric constraints imposed by their enclosing membranes. The cytoskeleton serves as a prime example, exhibiting sophisticated self-organisation that alters cell geometry in real time in response to physical stimuli.

DDC provides unprecedented flexibility, allowing us to apply instantaneous physical stimuli, either through direct contact or by modifying the local environment, with high spatiotemporal precision. Using this technique, we have conducted experiments across a diverse range of suspended cells, including microorganisms, yeast, and mammalian cells.

We investigated active turbulence in P. aeruginosa populations. Demonstrated cell sorting by separating Gram-negative (E. coli) and Gram-positive (M. luteus) bacteria from homogeneous mixtures. We applied geometric constraints and analysed how physical boundaries affect quorum sensing in E. coli. We also studied how circulating tumour cells survive the varying physical stresses of the circulatory system to successfully metastasise.

These studies were initiated in Türkiye, where the majority of the experimental work was completed prior to our relocation to Germany in 2023. Following the move, research involving live organisms was temporarily paused to secure Biosafety Level 1 (BSL-1) certification. With this accreditation successfully obtained in late 2025, we are in the final planning stages to resume laboratory operations. We are currently preparing active turbulence and cancer metastasis work for publication and expect to complete the remaining experimental work in the coming months.

Demonstration of (a) selective sorting of E. coli from M. luteus within an initially homogeneous mixed population, and (b) rapid collection and controlled relocation of S. cerevisiae cells using our DDC system.

We aim to take a step further by exploring analogues of “metabolism” and “evolutionary pathways” in colloidal aggregates of passive polystyrene spheres. In this context, metabolism refers to the continuous flux of particles, energy, and structure through an aggregate, while evolutionary pathways describe how these structures adapt, reorganise, or transition between states towards more stable solutions under driven conditions.

A central challenge is the stark mismatch in timescales. The intrinsic dynamics of the DDC system unfold on microsecond scales, whereas human observation and decision-making operate on timescales of seconds to minutes. As a result, the human operator can only access a coarse-grained view of the system: identifying where aggregates form, recognising stable pattern types, and noticing major perturbations such as the movement of a cavitation bubble boundary.

However, transient events, such as short-lived structures, rapid boundary fluctuations, local rearrangements, and subtle system-wide responses, remain largely inaccessible. In addition, there is an inherent latency in human intervention: observation, interpretation, decision, and action occur sequentially, introducing delays orders of magnitude slower than the dynamics being studied.

To overcome these limitations, we aim to develop an AI-driven operator that not only reproduces the qualitative observational capabilities of a human but also extends them through continuous, high-resolution quantification. Such a system can capture transient dynamics in real time, resolve fleeting configurations, and systematically track both local interactions and global system responses without the constraints of human perception or reaction time.

As a first step toward this goal, we start by developing methods to detect and track individual particles in full three-dimensional motion. This is non-trivial: although the DDC system is often treated as quasi-two-dimensional, particles frequently move out of plane, causing conventional tracking approaches to lose positional information. By combining tailored experimental design with machine learning algorithms, Dr.-Ing. Özgün Yavuz and Deniz Can Çağlar are working on establishing a framework that successfully reconstructs these 3D trajectories.

Sketch illustrating how a synthetic system's metabolism and evolution can be interpreted and analysed.

In addition to setting up our experimental capabilities to resume our previous work, we are currently designing an exciting research study on oligodendrocyte progenitor cells (OPCs) with Dr. Annika Haak. This will be a leap into understanding how the brain’s infrastructure uses physical force to navigate complex 3D environments. This is intended to be the first test of our new 3D tracking framework in a high-stakes biological context. Stay tuned!