New machine-learning method maps how bacteria become dangerous

Researchers in Hungary developed MALVINA, a machine-learning tool that tracks how bacteria invade human cells, damage DNA and respond to microbes and drugs. The method could help identify more dangerous strains and guide therapies that disarm pathogens instead of only killing them.

Why it matters: - MALVINA gives researchers a single-cell view of bacterial infection, showing not just whether bacteria survive but how they enter human cells, how they spread and how much DNA damage they cause. - The method could help distinguish harmless-looking infections from more dangerous ones that hide behind the same total bacteria count. - The approach may also support new drugs and microbiome-based strategies that reduce virulence instead of only killing bacteria.

What happened: - Researchers at the HUN-REN Biological Research Centre in Szeged and the Hungarian Centre of Excellence for Molecular Medicine developed MALVINA, short for Machine Learning-Based Virulence Interaction Analysis. - The method detects fluorescently labeled bacteria inside human cells and uses machine learning to analyze microscopy images. - The study was released on October 12, 2026, in Szeged, Hungary.

The details: - MALVINA measures bacterial entry into cells, accumulation inside cells and DNA damage in the same experimental system. - The tool can distinguish between many cells with a few bacteria and a few cells with many bacteria. - Bence Bognár, co-first author of the study, said the method lets researchers see infection “cell by cell” and separate those two infection patterns. - Szilvia Juhász, head of the Cancer Microbiome Research Group at HCEMM and corresponding author, said MALVINA shows how many bacteria entered the same host cell and what damage that cell suffered. - The researchers tested the method on four Escherichia coli strains, including a harmless laboratory strain, disease-associated strains that enter cells efficiently and strains that produce a DNA-damaging toxin. - The test revealed different virulence profiles across the strains. - One strain isolated from a colorectal tumor sample showed strong invasive capacity, intracellular accumulation and DNA damage under the tested conditions. - The team also tested how two bacterial strains behave together in the presence of human cells. - An invasive E. coli strain associated with inflammatory bowel disease helped an otherwise harmless laboratory strain enter host cells. - In other pairings, a genotoxic strain reduced the ability of non-genotoxic but invasive rivals to enter cells. - The researchers examined colibactin, a bacterial genotoxin produced by some gut-associated bacteria. - When colibactin production was disabled, the bacterium caused less DNA damage and was less able to suppress invasion by rival strains. - The team also tested antibiotics and other compounds that affect human cell functions. - Some treatments reduced the number of viable bacteria while increasing bacterial invasion into human cells.

Between the lines: - The findings challenge the idea that antibiotic success can be judged only by how many bacteria survive. - Bacterial virulence appears to depend on nearby microbes, host cells and drug exposure, not just on one strain acting alone. - That matters because standard averaged measurements can hide very different infection behaviors. - The work suggests some treatments may unintentionally reshape bacterial behavior even as they suppress growth. - A patent application has been filed for the MALVINA approach.

What's next: - The researchers say MALVINA could help identify especially invasive or DNA-damaging strains. - The method could also support the development of targeted microbiome-based or probiotic strategies. - Future drug research may use the approach to find treatments that disarm pathogens by reducing virulence, not just killing them. - The team says the platform can help determine whether reduced cell damage after treatment comes from fewer bacteria or from changed behavior in surviving bacteria.

The bottom line: - MALVINA turns bacterial infection into a measurable single-cell story, making it easier to see when microbes become dangerous and how drugs or neighboring bacteria change that risk.