Outline of the μFG. Credit: Science Advances (2026). DOI: 10.1126/sciadv.aea7598

Researchers led by Assoc. Prof. Dr. Savaş Taşoğlu from the Department of Mechanical Engineering at Koç University have developed a new, open-access and machine learning-assisted design tool aimed at automating microfluidic chip design. The research is published in Science Advances.

Microfluidic systems are chip-based platforms that enable biological and chemical samples to be processed in very small volumes with high precision and low cost, and they are finding increasingly widespread use. These systems are commonly employed in laboratories for cell culture studies, drug testing, diagnostic analyses, and experimental setups that mimic human tissues.

However, the design and fabrication of microfluidic chips often require advanced engineering expertise and numerous design iterations. This challenge stands out as one of the major obstacles faced by researchers who are new to working with microfluidic technologies.

Introducing the μFluidicGenius tool

To address this challenge, researchers from Koç University developed a new microfluidic chip design tool called μFluidicGenius (μFG). The tool is intended to enable especially nonexpert users to rapidly design microfluidic circuits that meet their target flow conditions.

The core approach of μFluidicGenius is to make microfluidic chip design more accessible in a software environment by relying on limited and intuitive user inputs. Users define the placement of reservoirs on the chip within a predefined area, specify the channel connections between them, and enter the target flow rates desired for each section. In this way, the design process can proceed without requiring users to directly engage with complex fluid mechanics calculations.

μFG employs a hybrid approach that combines machine learning models with mathematical fluid mechanics calculations to determine the fluidic resistance values required to achieve the desired flow distribution. To realize these resistance values, the system generates maze-like channel geometries within the chip that can provide precisely tuned resistance characteristics. This allows the required flow conditions to be optimized within a limited chip footprint.

Supporting complex physiological flow profiles

The study also shows that the developed method is not limited to simple flow configurations. μFluidicGenius enables the design of complex flow profiles, such as those required for multi-organ-on-chip platforms that resemble physiological conditions. This capability makes it possible to address microfluidic systems with diverse experimental requirements using the same design approach.

Microfluidic circuits created with μFluidicGenius can generate output files that are directly compatible with three-dimensional (3D) printing. Experimental results indicate that, in chips designed with μFG and fabricated using 3D printing, the measured flow rates match the target values with approximately 90% accuracy. The study also demonstrates that the system can successfully accommodate both low and high fluidic resistance requirements within a fixed and limited chip area.

Publication details

Mehmet Tugrul Birtek et al, ML-automated microfluidic circuit design, Science Advances (2026). DOI: 10.1126/sciadv.aea7598

Citation: Automating microfluidic chip design: Hybrid approach combines machine learning with fluid mechanics (2026, February 2) retrieved 2 February 2026 from https://phys.org/news/2026-02-automating-microfluidic-chip-hybrid-approach.html

This document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, no part may be reproduced without the written permission. The content is provided for information purposes only.