← Latest papers
💻 bioinformatics

DNA methylation assigns sequencing reads to closely related bacterial strains when their sequence cannot

The paper introduces Epilect, a method that leverages strain-specific DNA methylation patterns from nanopore sequencing to accurately assign reads to closely related bacterial strains that are indistinguishable by sequence alone, thereby enhancing strain-level resolution in clinical and public health contexts.

Original authors: Kakabra, K., Claesson, M. J.

Published 2026-11-15
📖 5 min read🧠 Deep dive

Original authors: Kakabra, K., Claesson, M. J.

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ⚕️ This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer

In the microscopic world of bacteria, the difference between a harmless resident and a dangerous invader often comes down to the specific strain, not just the species. Imagine two bacteria that look identical under a microscope and share nearly the same genetic code; yet, one might carry a deadly antibiotic resistance gene while the other does not. For scientists trying to track disease outbreaks or understand how infections persist in a patient, identifying which specific strain is present is crucial. The standard tool for this is DNA sequencing, which reads the genetic letters of an organism to identify it. However, when two strains are extremely closely related, their genetic letters are so similar that a single long strand of DNA often looks exactly the same in both. In these cases, the genetic code alone cannot tell the story of where that strand came from, leaving researchers with a blind spot in their surveillance.

Bacteria, however, carry a second layer of information hidden within their DNA that is invisible to standard genetic reading. This is a chemical tag called methylation, which acts like a molecular stamp placed on specific patterns of genetic letters. Different bacterial strains often use different stamps, marking different patterns, even if their underlying genetic code is nearly identical. A newer type of sequencing technology, known as nanopore sequencing, has the unique ability to read these chemical stamps directly as it reads the genetic letters, without needing any extra chemical processing. This capability opens a new door for distinguishing between bacterial twins that genetic sequencing alone cannot separate.

A team of researchers set out to measure just how often this genetic blind spot occurs and whether these chemical stamps could provide the missing clues. They analyzed a large collection of bacterial samples from four common species, including those that cause serious infections in hospitals. By comparing pairs of strains that were nearly identical, they discovered a significant limitation in relying solely on genetic sequence. In pairs of strains that shared more than 99 percent of their genetic code, nearly half of the DNA strands read carried absolutely no genetic difference between the two. When the strains were even more similar, sharing over 99.9 percent of their code, about three-quarters of the DNA strands were indistinguishable by their genetic sequence alone. This finding confirmed that for very closely related bacteria, the genetic map often fails to provide a unique address for every piece of DNA.

To solve this, the researchers developed a new method called Epilect, which uses the chemical stamps to assign each DNA strand to its correct strain. Instead of looking for differences in the genetic letters, the software looks for differences in the chemical patterns. They tested this approach on a controlled mixture of two bacterial strains that had different chemical stamps but nearly identical genetic codes. The results showed that while many individual DNA strands carried very little information, the method could still correctly identify the source of the strand with high accuracy. When the researchers applied this to the larger collection of bacterial pairs, the method successfully assigned about 22 percent of the previously indistinguishable DNA strands to the correct strain. Crucially, it did this with a very low error rate, meaning that when it made a call, it was almost always right.

The study also demonstrated that scientists can predict whether this method will work for a specific pair of bacteria before they even begin sequencing. By simply looking at the lists of chemical patterns each strain is known to mark, researchers can determine if the two strains have different enough stamps to be told apart. The researchers found that this prediction was accurate for the vast majority of the bacterial pairs they tested. In situations where the genetic code was too similar to provide clues, but the chemical stamps were different, this new method outperformed existing tools that rely only on genetic sequence. However, the method is not a universal replacement; it works best when the strains are very similar and have distinct chemical patterns. If the strains share the same chemical stamps, the method correctly declines to make a guess, acknowledging that the information is not there.

This work provides a practical way forward for laboratories using modern sequencing technology. It shows that for closely related bacterial strains, the chemical stamps on the DNA offer a second, independent source of evidence that can recover information lost by the genetic sequence. The researchers made their software and data publicly available, allowing other scientists to check in advance whether their specific bacterial samples can be distinguished by these chemical patterns. This means that in the future, when tracking the spread of infections or studying how bacteria evolve inside a patient, scientists will have a more complete picture, able to see the subtle differences between bacterial twins that were previously invisible.

Drowning in papers in your field?

Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.

Try Digest →