Biochip uses AI to identify genetic markers in 20 minutes

NTU Singapore researchers built an AI-powered biochip that detects disease-linked microRNAs in about 20 minutes.

Researchers at Nanyang Technological University, Singapore have built a biochip that can spot disease-linked genetic markers in about 20 minutes, a task that often takes conventional lab methods much longer. The device targets microRNAs, tiny molecules that help control gene activity and can serve as early warning signs for illnesses such as cancer and heart disease. Instead of relying on polymerase chain reaction, or PCR—a common technique that amplifies genetic material through multiple processing steps—the new system reads signals directly from a small sample. It combines a nanophotonic chip, which manipulates light at a very small scale, with artificial intelligence that interprets the resulting image in one pass. That design lets the platform capture and analyze thousands of signals in a single snapshot rather than one by one. In tests, the researchers used it to measure three microRNAs associated with non-small cell lung cancer and showed that it could work in biologically realistic samples. The study, published in Advanced Materials, points to a faster and potentially simpler route to molecular testing that could one day fit into more accessible clinical or point-of-care tools.

How the chip works

A useful way to picture the device is to imagine a dark room where even a faint glow becomes obvious. The NTU team created a light-trapping structure called a nanocavity that intensifies fluorescent signals, making very small molecular events much easier to see.

Those signals appear when a probe on the chip binds to its target microRNA. Because the optical design strengthens that glow, the system can detect extremely low amounts of material and even distinguish between microRNAs with very similar sequences, down to single-molecule sensitivity according to the source description.

Why microRNAs matter

MicroRNAs are short strands of genetic material that do not code for proteins but instead regulate how genes are switched on or off. That makes them biologically important and medically useful: changes in their levels can reflect disease processes well before symptoms become obvious.

Researchers have been interested in using microRNAs as biomarkers, measurable biological signs of disease, for years. The challenge is that these molecules are small, often scarce, and difficult to separate cleanly from very similar genetic sequences, which is why testing can be slow and technically demanding.

What AI adds

The chip's optical hardware is only part of the story. The other key piece is AI-driven image analysis, which helps the system interpret a complex field of signals captured all at once.

Think of it like looking at a crowded night skyline and having software instantly tell you which lights matter. Here, the AI analyzes the image produced by the chip and identifies patterns linked to different microRNA targets, allowing thousands of signals to be processed from a single snapshot rather than through a longer sequence of manual or instrument-heavy steps.

What the researchers tested

According to the source, the team successfully measured three microRNAs tied to non-small cell lung cancer: miR-191, miR-25, and miR-130a. Importantly, they did this directly from human lung cancer cell extracts, which is a more realistic test than working only with idealized samples.

The researchers also added synthetic microRNAs to biological samples and reported that the chip still performed reliably. That matters because many diagnostic ideas work under controlled lab conditions but struggle once real biological material introduces background noise and chemical complexity.

Designed for speed and access

The prototype was built with usability in mind, not just raw analytical performance. The team incorporated a color camera and a mobile app, suggesting that the platform is being shaped as a practical tool rather than a bench-top demonstration that only specialists could operate.

Using only a tiny drop of blood, the system is designed to detect multiple biomarkers without amplification or complicated preparation steps. Cutting out those extra stages could reduce both the time and the expertise needed to run a test, which is often just as important as sensitivity when a technology moves toward real healthcare settings.

Why This Matters

Fast molecular testing matters because timing changes what clinicians can do. If a doctor can identify a meaningful pattern of biomarkers quickly, treatment decisions, follow-up testing, and patient monitoring can happen sooner and with less friction.

This study also speaks to a larger shift in diagnostics: combining advanced sensing hardware with software that can make sense of dense biological data. That pairing could help move sophisticated genetic analysis out of specialized labs and into settings where speed, simplicity, and lower sample volume make a real difference.

What comes next

Associate Professor Chen Yu-Cheng, who led the work at NTU's School of Electrical and Electronic Engineering, said the longer-term vision is an automated system that could use blood or saliva to screen for hundreds or even thousands of biomarkers at once. That would be a big step toward broader screening and more personalized medicine, where treatment or monitoring is tailored to an individual's biology.

There is still a gap between a promising prototype and routine clinical use, and the source does not claim that the device is there yet. But by showing direct microRNA detection in minutes, with a design aimed at multiplexing many targets at once, the NTU team has outlined a practical direction for faster and more accessible genetic testing.