🦠 microbiology

A membrane-impermeant nucleic acid dye converts bacteriophage plaque assays into a machine-readable format for automated counting

This study demonstrates that using a membrane-impermeant nucleic acid dye to fluorescently label bacteriophage plaques creates high-contrast images compatible with simple, open-source automated counting pipelines, thereby enabling high-throughput, statistically rigorous quantification across diverse phage-host systems without requiring genetic engineering or complex machine learning.

Original authors: Wiwi, A., Arnold, J., Branch, D., CAHILL, J.

Published 2026-08-09

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Wiwi, A., Arnold, J., Branch, D., CAHILL, 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 ⚕️ This is an AI-generated explanation of a preprint that has not been peer-reviewed.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

Imagine the microscopic world of viruses as a high-stakes game of hide-and-seek played on a giant, jello-like stage. In this game, the players are bacteriophages (or "phages" for short), which are tiny viruses that hunt bacteria. To see who wins, scientists use a classic trick called a "plaque assay." They mix the phages with a swarm of bacteria in a soft, gelatinous layer. Wherever a phage finds a bacterium and bursts it open, it leaves behind a tiny, invisible crater. Over time, these craters grow into clear, round spots called "plaques," looking like little moons on a starry night. Counting these moons tells scientists how many phage viruses are in a sample.

However, there's a catch. These "moons" are often faint, fuzzy, or blend right into the background, making them incredibly hard to count by eye. It's like trying to count snowflakes on a cloudy day. For decades, scientists have had to squint at petri dishes for hours, manually tallying every single spot. This is slow, boring, and prone to human error, especially when you need to run hundreds of experiments at once. The big question has been: How can we make these invisible spots glow so bright that a computer can count them instantly, without needing to genetically engineer the viruses or build a super-expensive robot?

This paper introduces a clever, low-tech solution that turns these invisible craters into glowing beacons. The researchers used a special dye called SYTOX, which acts like a "glow-in-the-dark" paint that only sticks to dead things. Normally, this dye can't get inside healthy, living bacteria because their cell walls are like strong, sealed doors. But when a phage attacks and bursts a bacterium, it breaks that door. The dye rushes in, hugs the exposed DNA inside, and suddenly, the dead spot lights up with a brilliant fluorescent glow.

The team tested this "glow-up" method on a variety of phage viruses, including some that are round, some that look like long worms, and some that are even wrapped in a bubble-like envelope. They found that for most of these viruses, adding the dye to the gel before the experiment started made the plaques pop out in high-definition fluorescence. When they took pictures of these glowing plates and ran them through free, open-source software (a program called ImageJ), the computer could count the spots almost perfectly, matching the counts of human experts. It was as if the computer suddenly got super-vision, seeing the plaques clearly where the human eye struggled.

The method wasn't perfect for every single scenario, though. The researchers discovered that for one specific type of phage (M13), which doesn't kill its host immediately but instead slowly leaks out new viruses, the glow didn't appear until the next day. This actually makes sense biologically, as the cells stay alive longer, keeping the dye out until they finally give up. They also found that for one tough, Gram-positive bacteria system, adding the dye before the experiment actually stopped the plaques from forming at all. But, they had a backup plan: they simply waited for the plaques to form first, and then added the dye afterward. This "post-party" approach worked just as well, proving the method is flexible.

Perhaps most importantly, the scientists checked if this glowing dye would ruin the viruses for future use. They picked some of the glowing spots out of the gel and tried to grow them again. For most viruses, the dye didn't hurt them at all. For one virus (MS2), the dye did slightly reduce the number of survivors, but it wasn't a deal-breaker. This suggests that scientists can use this glowing trick to count their viruses quickly and accurately, and then still use those same viruses for further experiments.

In short, this paper shows that you don't need fancy robots or genetic engineering to automate virus counting. You just need a little bit of glow-in-the-dark dye and a free computer program. By turning faint, hard-to-see spots into bright, machine-readable signals, this simple trick could help scientists run bigger, more precise experiments faster, turning the tedious job of counting viral craters into a quick, automated task.

Technical Summary: Membrane-Impermeant Nucleic Acid Dye for Automated Bacteriophage Plaque Counting

Problem Statement
Plaque assays remain the gold standard for quantifying infectious bacteriophage particles, yet they suffer from significant practical limitations. Manual counting is labor-intensive, time-consuming, and difficult to scale for large experiments involving multiple strains, dilutions, and replicates. Furthermore, the precision of these assays is fundamentally constrained by Poisson counting statistics; to achieve higher precision, researchers must count more plaques, which exacerbates the manual burden. While automated image analysis tools exist, their performance is often limited by the low contrast and heterogeneous morphology of conventional white-light plaque images, particularly for faint, small, or turbid plaques. Current solutions often require phage engineering, specialized machine-learning pipelines, or custom imaging hardware, making them inaccessible to many standard laboratories.

Methodology
The authors developed a workflow to convert standard soft-agar overlay plaque assays into a high-contrast, machine-readable format using a membrane-impermeant nucleic acid dye (SYTOX). The core hypothesis is that the dye, which is excluded from intact cells but fluoresces upon accessing nucleic acids following membrane compromise or lysis, will generate a strong fluorescent signal specifically within plaque zones.

The study evaluated this approach across diverse phage-host systems, including:

Lytic dsDNA phages: T7 (Escherichia coli) and Leo2 (Bacillus safensis).
Lytic ssRNA phage: MS2 (E. coli).
Filamentous ssDNA phage: M13 (E. coli).
Enveloped dsRNA phage: phi6 (Pseudomonas syringae).

Two primary workflows were tested:

Pre-incorporation: SYTOX was added directly to the molten soft agar before plating.
Post-labeling: For systems where pre-incorporation inhibited plaque formation (specifically Leo2), SYTOX was applied to the agar surface after plaque development.

Imaging was performed using a standard Syngene G:BOX Chemi XX6 system with Alexa Fluor 500 settings. Automated counting was executed using the open-source Fiji/ImageJ software suite, specifically utilizing the "Find Maxima" function. The analysis pipeline required no custom software or machine learning, relying instead on adjusting "prominence" parameters to distinguish plaque-associated maxima from background noise. The method's impact on downstream applications was assessed by picking individual plaques from labeled plates and re-titering them to check for loss of infectivity.

Key Results

Enhanced Contrast and Automation: In compatible systems (T7, MS2, phi6), fluorescent labeling produced high-contrast images where plaques appeared as discrete fluorescent puncta. Automated counts using the "Find Maxima" algorithm agreed closely with manual counts, falling within the expected Poisson uncertainty ( $\pm\sqrt{N}$ ).
Early Detection: Fluorescent signals often emerged prior to or alongside the visibility of conventional plaques, allowing for machine-readable counting during early stages of plaque development.
Diverse Applicability: The method successfully automated counting for phages with distinct genome types (dsDNA, ssRNA, ssDNA, dsRNA) and lifestyles (lytic vs. chronic).
- M13 (Chronic): As expected for a non-lytic phage, M13 did not show useful fluorescence on the same day. However, by the conventional 24-hour endpoint, sufficient membrane compromise occurred to allow fluorescent detection and automated counting.
- T7 (Large Plaques): While early-stage T7 plaques were easily counted, large, mature plaques (24h+) exhibited peripheral dye accumulation (ring-like labeling), which initially hindered counting. This was resolved by using the post-labeling workflow or adjusting analysis parameters.
Compatibility Limits and Solutions: In the Gram-positive Bacillus safensis system (Leo2), pre-incorporation of the dye inhibited plaque formation entirely, despite normal lawn growth. A post-labeling workflow (applying dye after plaque formation) successfully restored fluorescent labeling and enabled accurate automated counting.
Downstream Viability: Plaque recovery experiments indicated that fluorescent labeling generally preserved phage infectivity. While MS2 showed a statistically significant reduction in recoverable infectivity after dye exposure, this did not measurably impair plaque formation in the overlay assay. Leo2 plaques recovered via the post-labeling workflow showed no detectable loss of viability compared to controls.

Significance and Claims
The paper claims to provide a practical, accessible route to higher-throughput and statistically rigorous phage biology without requiring phage engineering, specialized imaging infrastructure, or complex machine-learning pipelines. By converting standard plaque assays into a high-contrast format compatible with simple, open-source software (ImageJ), the method addresses the primary bottleneck of manual counting.

The authors emphasize that the technique does not necessarily outperform human counters in absolute terms but renders plaque assays sufficiently high-contrast that simple software can perform the task reliably. The study also highlights that the timing and morphology of the fluorescent signal reflect underlying biological differences (e.g., lytic vs. chronic infection, plaque maturation dynamics), suggesting potential utility as a phenotypic indicator. Ultimately, the work aims to lower the barrier to entry for automated phage quantification, supporting larger experiments with more replicates and stronger statistical confidence in both low-resource and automation-oriented laboratories.

More like this