Scientists use Light and AI to measure DNA fragment length for the first time

Scientists have shown that the length of DNA molecules can be measured using light and artificial intelligence (AI), opening up a potential new way to analyse genetic material. In the future, this could make important cancer screening tests more accessible around the world.

Led by researchers at The Institute of Cancer Research, London, and the Optoelectronics Research Centre at the University of Southampton, the study combines machine learning with vibrational spectroscopy – a technique that uses light to analyse molecules. The approach allows researchers to estimate DNA fragment length without relying on expensive laboratory equipment or physically separating DNA molecules, as is required with conventional methods.

The findings, published in ACS publications (full links below), establish a new approach to DNA analysis and highlight the power of combining biological and engineering sciences to tackle long-standing challenges in genomic research.

Funding was predominantly provided by the University of Southampton, including an EPSRC DTP studentship for Rashad Fatayer, who led the experimental/AI work with additional support from The Institute of Cancer Research – which is both a research institute and charity – and Breast Cancer Now.

Why DNA length matters

The length of DNA fragments is an important measurement in many areas of molecular biology. It is routinely assessed in genomic experiments and can provide valuable insight into the health and behaviour of cells.

In cancer research, DNA fragment length is attracting growing interest as a biomarker – an objectively measurable indicator of disease. Small fragments of DNA released from tumours circulate in the bloodstream as circulating tumour DNA (ctDNA) and are typically shorter than those released from healthy cells. Analysing fragment length can help monitor disease progression and assess how patients are responding to treatment.

Despite its importance, measuring DNA fragment length remains technically demanding. Researchers commonly use gel electrophoresis-based systems to physically separate DNA by size, while sequencing-based approaches are often used for analysing ctDNA. These methods are highly effective but require specialised equipment, can be expensive and may consume or alter the sample – limiting their accessibility, particularly in lower-resource settings.

Using light to analyse DNA

To address this challenge, the research team investigated whether DNA length could be determined using vibrational spectroscopy.

This technique works by shining light onto a sample and measuring how the molecules interact with it, which generates unique spectral patterns, or molecular ‘fingerprints’, that reflect their structure and composition.

While spectroscopy has previously been used to study biological materials, this is the first time it has been shown to estimate DNA fragment length directly from these optical signatures.   

The team combined spectroscopic measurements with machine learning algorithms trained to detect subtle patterns linked to DNA fragments of different lengths. Using this approach, the models were able to estimate DNA fragment length without physically separating the molecules.

Importantly, the approach requires only a very small amount of sample, does not destroy the DNA and could allow samples to be recovered for further analysis. The results demonstrate proof of principle that DNA fragment length can be predicted from spectroscopic data alone.

The work builds on a collaboration between researchers at The Institute of Cancer Research (ICR) and the University of Southampton, with the team filing a patent application covering the technology, underlining its potential for future development.

Adjacently, the team worked on another study, also published in bioRxiv, describing the use of the spectroscopy technology to measure DNA methylation – a genetic modification that regulates gene expression, and when disrupted, is associated with several human diseases, including cancer. 

Future steps

The researchers now plan to build on these findings by exploring how the approach performs in more complex biological samples and by investigating its potential applications in ctDNA analysis.

Study author Dr Stephen-John Sammut, Group Leader of the Cancer Dynamics Group at the ICR, said: “Further studies will be required to improve the accuracy of the models, validate the approach against established laboratory methods and determine how it might be integrated into future genomic workflows.

“While the research will not have an immediate impact on patients, it opens up a completely new avenue of investigation. By showing that spectroscopy and AI can be used to estimate DNA fragment length for the first time, the study lays the groundwork for future technologies that could make important genetic analyses faster, cheaper and more accessible around the world.”

Lead author Professor Senthil Murugan Ganapathy, Head of the Integrated Photonic Devices Group at the Optoelectronics Research Centre, said: “Light interacts with biological molecules almost instantaneously, producing unique molecular signatures that reveal important information about their structure and composition. This study shows that, when combined with AI, vibrational spectroscopy can reveal characteristics of DNA that previously required complex laboratory techniques.

“Because spectroscopy is rapid, non-destructive and requires only simple instrumentation and very small sample volumes, it has enormous potential to bring advanced molecular analysis much closer to patients. As the technology matures, it could become an indispensable bedside tool, helping clinicians detect disease earlier, monitor treatment more effectively and deliver faster, more accessible diagnostics.”

Read both papers in full via the links below

DNA Fragment Length Analysis Using Machine Learning Assisted Vibrational Spectroscopy

DNA Methylation and Hydroxymethylation Quantification Using Vibrational Spectroscopy