Seminar on Programmable 3DP microfluidic bio-reaction system: automated LAMP-on-a-chip

Koç University, Turkey

This seminar will be chaired by A/P Huang Changjin.

Microfluidics enables high-precision, cost-effective processing of biological and chemical substances, powering applications from lab-on-chip platforms to point-of-care (PoC) diagnostics. However, designing and fabricating functional microfluidic circuits traditionally demands substantial expertise, specialized software, and iterative prototyping, creating significant barriers for non-experts. To address this, we introduce µFluidicGenius (µFG), an open-access, machine learning (ML)-augmented design tool that democratizes the creation of complex microfluidic systems. Users simply specify the spatial placement of reservoirs, define channel connections, and assign target flow rates; µFG then employs a hybrid framework combining ML models with mathematical modeling to automatically generate spatially optimized maze-like resistive structures that achieve the precise fluidic resistances required for the desired flow distribution. These resistive elements are efficiently packed within user-defined geometries and can implement sophisticated flow profiles, including physiologically relevant conditions for multi-organ-on-chip or PoC devices. Designs are directly exportable for 3D printing, and experimental validation confirms that µFG-generated circuits reproduce target flow distributions with approximately 90% accuracy.

By automating and simplifying microfluidic circuit design, µFG lowers the entry barrier for non-specialists while exemplifying a principled integration of machine learning into fluidic engineering for rapid, customizable device development. As a practical demonstration of its potential, µFG facilitates the engineering of advanced PoC platforms, such as those featuring three-dimensional hydrophobic valves that serve as programmable bio-reaction reservoirs. These valves, realized via 3D-printed soft lithography, offer tunable burst pressures (ranging from 6.4 to 44.8 mbar) and support both series and parallel configurations for controlled fluid handling across diverse samples (e.g., water, blood, serum). When paired with a portable pressure pump, they enable precise flow programming. Notably, this approach has been applied to integrate consecutive bio-reservoirs— with lyophilized primers—into a compact microfluidic chip for automated loop-mediated isothermal amplification (LAMP) detection of the Mpox virus, achieving visible colorimetric readout suitable for field-deployable diagnostics. Together, these advances highlight how ML-driven tools like µFG accelerate the translation of innovative microfluidic concepts into functional, impactful PoC technologies.