AI-powered biochip detects genetic markers in 20 minutes
Scientists at NTU Singapore developed an AI-powered biochip that detects multiple microRNA biomarkers in 20 minutes. The platform achieved more than 99 per cent accuracy in test channels.
A team of scientists from NTU Singapore has developed a new biochip that, when paired with Artificial Intelligence (AI), can quickly and accurately detect extremely small amounts of microRNAs, tiny genetic markers linked to diseases such as heart disease. Published in Advanced Materials, the new biosensing platform combines a specially designed nanophotonic chip with AI-automated image analysis. Compared with the current gold standard of detecting microRNA – PCR (polymerase chain reaction) – the new device can cut detection time from hours to 20 minutes.
With a tiny drop of blood loaded into the chip, it can rapidly detect multiple microRNA biomarkers. With its integrated AI imaging function, thousands of microRNA signals can be imaged and analysed in a single snapshot. The team has constructed a compact prototype that includes a colour camera that can capture images of the nanophotonic chip, and a mobile phone application designed to analyse images for microRNA using AI algorithms and provide rapid results.
MicroRNAs are short RNA molecules that help regulate genes that work in the body. Because changes in microRNA levels are linked to many diseases, scientists have been studying them as possible biomarkers for conditions such as cardiovascular disease, cancer, neurodegenerative disorders and metabolic illnesses. The importance of microRNAs was underscored in 2024, when the Nobel Prize in Physiology or Medicine recognised the discovery of microRNA and its role in gene regulation.
To overcome the challenges of detecting microRNA, the team designed a nanocavity, a tiny light-trapping structure hundreds of times smaller than the width of a human hair. Shaped like a cave lined with mirrors, the nanocavity reflects and boosts fluorescent signals that glow when a target microRNA binds to its matching probe. This makes it easier to detect even single microRNA molecules.
The system measured three microRNAs associated with non-small cell lung cancer — miR-191, miR-25, and miR-130a — from human lung cancer cell extracts without amplification or complex preparation. Unlike PCR and hybridisation kits that require labelled probes, the platform directly and quantitatively detects multiple microRNAs in liquid samples. The platform also uses a deep-learning model known as Mask R-CNN to automatically analyse microscopic images.
An automated AI imaging system captures the microRNA signals in one shot, after which the AI system identifies and classifies fluorescent signals and distinguishes between different microRNA types, removing the need for manual counting and reducing human error. The platform also performed well when synthetic microRNAs were added to biological extracts, suggesting that it can work reliably in more realistic sample conditions. The researchers said the platform could detect microRNAs at extremely low concentrations, down to just a few molecules in a sample, and achieved more than 99 per cent accuracy in identifying its targets across different test channels.
The team aims to build a system that can quickly and accurately measure multiple microRNAs, with the potential to detect biomarkers linked to a wide range of diseases. Their successful tests with lung cancer cells show that, with the right probes targeting different biomarkers, the technology could potentially be adapted for many other cancers and diseases, including cardiovascular and viral diseases. A technology disclosure has been filed through NTUitive, the university’s innovation and enterprise company.