Biology constantly challenges and explores the limits of what defines an independent organism. Just when the scientific community settles on a functional definition of cellular life, the natural world reveals an entity that shatters those parameters.

A recent identification of an unprecedented marine microbe found in Japan has done exactly that. The microorganism (or is it?) is the archaeon Candidatus Sukunaarcheum mirabile isolated from a dinoflagellate (single-celled eukaryotic plankton).

The archaeon appears to blur the line between a free-living cell and a parasitic virus, offering an extraordinary opportunity to re-evaluate one of molecular evolution's most enduring mysteries: the origins of viruses.

Candidatus Sukanarcheum mirabile holds the key to understanding how complex biological systems can strip themselves down to their barest essentials.

The microorganism's extreme genome reduction challenges established frameworks and demands a rigorous re-examination of how parasitic entities evolve in the natural world.

Background on the Discovery

Isolation from a Marine Host

Researchers identified Candidatus Sukanarcheum mirabile using advanced single-cell genomic amplification techniques targeting a marine dinoflagellate known as Citharistes regius.

Researchers collected these dinoflagellates off the coast of Shimoda, Japan, using a plankton net with a 25-micrometre mesh.

After sequencing the amplified DNA from the microbial community residing within and on this host, scientists detected a highly unusual circular sequence, which was later confirmed to be widely distributed across marine environments, notably in Brazil.

Interestingly, these related sequences were found mainly in the 5 to 20 micrometre size fraction of plankton. Given that typical free-living archaea are about 1 micrometre in diameter, this size distribution suggests these archaea likely live in close association with larger eukaryotic host cells rather than floating freely in the ocean.

A Minimal Genome

The genome of Sukunaarchaeum is remarkably small, measuring just 238 kilobase pairs, representing the smallest complete archaeal genome ever recorded.

It is less than half the size of the previous archaeal record holder, Nanoarchaeum equitans. The genome of Sukunaarchaeum possesses 189 protein-coding genes.

To put this staggering genetic minimisation into perspective, human cells encode roughly 20,000 proteins, and even highly reduced symbiotic bacteria typically retain many hundreds of genes.

The sheer capacity of genetic material is fascinating, yet the true marvel lies in which genes the organism chose to keep and which it systematically discarded over millions of years of evolution.

Metabolically Helpless

Metabolic pathways are absent from the genome of Sukunaarchaeum. The organism encodes absolutely no recognisable mechanisms for generating its own energy, synthesising amino acids, producing vitamins or building carbohydrates. It is completely metabolically helpless.

A typical bacterial endosymbiont, as well as ancient endosymbiotic organelles (eg mitochondria and chloroplasts), usually retains specific biosynthetic pathways to produce compounds beneficial to its host.

For example, the highly reduced bacterium Candidatus Carsonella ruddii lives inside sap-feeding insects and has a tiny genome of 16o kilobase pairs. However, it still retains genes dedicated to producing amino acids for its host.

Sukanaarchaeum offers no such mutualistic benefit, strongly suggesting an obligate parasitic lifestyle where it unilaterally exploits external resources for survival. It almost passed to be a virus if it were not stubbornly holding onto its ribosomes.

A Replicative Core in Isolation

Despite this profound metabolic depletion, the archaeon retains a complete replicative core. It holds onto the genetic blueprints for ribosomes, transfer RNAs, DNA polymerases and the essential machinery required for transcription and translation.

Over half of its protein-coding genes are dedicated to genetic processing, creating a bizarre biological paradox. The microorganism essentially functions as a cellular entity entirely dedicated to self-replication.

It relies on the host for every raw material and unit of energy it needs to exist, while focusing its limited genomic real estate strictly on propagating its own genetic information.

To fully grasp the significance of this discovery, it is essential to understand the sheer scale of the unknown proteins encoded within this tiny genome.

Despite facing immense evolutionary pressure to shrink its genetic footprint, Sukunaarchaeum dedicates an astonishing 25 per cent of its entire genome to just eight unusually large proteins. Thirteen of its proteins exceed 1 000 amino acids in length, with the largest reaching 4 756 amino acids.

These massive proteins share no primary sequence homology with any known cellular proteins. Structural prediction algorithms suggest they are packed with transmembrane helices, the kind that associate with the cell membrane.

This massive investment in membrane proteins while discarding basic metabolism is highly uncharacteristic for cellular life, providing a vital clue to how this organism interacts with its environment.

The Three Hypotheses of Viral Origins

The discovery of such an extremely reduced cell naturally feeds into the long-standing debate regarding how viruses first emerged on Earth. Because viruses do not leave physical fossil records in sedimentary rock, scientists must painstakingly reconstruct their origins using comparative genomics and structural biology.

The scientific community generally categorises viral origins into 3 primary hypotheses. Each model provides a distinct framework for understanding how parasitic genetic elements came to dominate the biosphere.

1. The Virus-First Hypothesis

The virus-first hypothesis suggests that viruses actually predate modern cellular life. According to this framework, viruses originated in a pre-cellular RNA world as independent, self-replicating molecular units.

As the primordial soup gradually gave rise to complex, membrane-bound cells, these ancient viral replicators adapted to parasitise them. Proponents of this theory argue that the immense genetic diversity found in the virosphere points to an ancient origin that stretches back billions of years, well before the Last Universal Common Ancestor (LUCA) of modern cells.

The existence of viral hallmark genes that have no homologues in cellular life is often cited as evidence that viruses possess a fundamentally independent lineage.

While the concept of ancient replicators is mathematically sound, the virus-first hypothesis suffers from a severe logical contradiction. Modern viruses are obligate parasites.

Viruses absolutely require a host cell to provide energy and translational machinery to reproduce. It is exceedingly difficult to rationalise how a strictly dependent parasite could evolve and thrive before the existence of the very host it relies upon.

If primordial viruses were completely autonomous entities capable of self-replication without a host, they would essentially be classified as primitive cells rather than true viruses.

The transition from a free-living ancient replicator to a modern obligate parasite remains a significant conceptual gap in this specific model. Some theorists attempt to bridge this gap by suggesting viruses co-evolved alongside ancient proto-cells, but the fundamental contradiction of an autonomous parasite remains problematic.

2. The Escape Hypothesis

Often referred to as the progressive theory, the escape hypothesis posits that viruses are rogue genetic elements that gained the ability to exit one cell and enter another.

Plasmids and transposons are mobile genetic elements known to jump between regions of the host genome or transfer horizontally between adjacent cells.

The escape hypothesis argues that some of these selfish genetic elements occasionally acquired structural proteins, allowing them to form protective capsids and leave the host environment entirely.

This concept was heavily championed in the late twentieth century, with researchers suggesting that retroviruses evolved from cellular components that slipped out of regulatory control.

This progressive model neatly explains the origin of many small, simple viruses. It provides a logical mechanism for how bacteriophages might have evolved from escaping prokaryotic genes or how specific human viruses might have splintered off from eukaryotic cells.

The presence of proviruses integrated into modern cellular genomes is frequently highlighted as evidence of this genetic fluidity.

However, the escape hypothesis encounters severe limitations when confronted with the immense complexity of nucleocytoplasmic large DNA viruses.

Giant viruses such as Mimivirus and Pandoravirus boast massive genomes that rival those of parasitic bacteria. These giant viruses encode hundreds of complex proteins, many of which are unique to the viral world.

If viruses simply escaped from cellular hosts, one would expect their genomes to resemble a subset of host DNA. The presence of highly specific viral hallmark genes and massive bespoke protein structures strongly suggests that giant viruses did not simply break away from modern host cells as rogue plasmids.

The escape theory struggles to account for the sophisticated transcription machinery and complex morphogenetic networks found in these giant viral lineages.

3. The Reduction Hypothesis

The reductionist, or regressive, hypothesis presents a completely different evolutionary trajectory. This model argues that viruses were once ancient, autonomous cells.

Over vast stretches of evolutionary time, these free-living organisms adopted parasitic replication strategies. Because the host cell provided a stable environment rich in nutrients and energy, the parasite no longer needed to maintain its own metabolic pathways.

Genes that were no longer subjected to positive selection pressure gradually degraded and were permanently lost.

Eventually, the organism discarded its entire metabolic apparatus, losing the ability to independently synthesise proteins and crossing the threshold into what we now classify as a virus.

This hypothesis elegantly explains the sheer size and complexity of giant viruses. According to the reductionist view, giant viruses represent the intermediate stages of genomic decay.

They are the shrinking remnants of an ancient cellular domain that chose the path of extreme parasitism.

Proponents of this theory argue that giant viruses retain vestigial components of their cellular past, such as the ability to encode their own transfer RNAs and complex DNA repair enzymes.

Until recently, a major criticism of this theory was the lack of transitional cellular forms in the modern environment. Critics argued that the leap from a complex parasitic cell to a minimal virus was too vast without a clearly identifiable missing link.

While endosymbiotic bacteria like Mycoplasma demonstrate genome reduction, they still retain significant metabolic autonomy compared to a virus.

The scientific community needed an example of a cellular organism that had pushed reduction to the absolute limit without losing its ribosomes.

The Case for the Reductionist Theory

The unearthing of Sukunaarchaeum provides an extraordinary piece of empirical evidence that reinvigorates the reductionist hypothesis.

This deeply branching archaeon functions as a modern blueprint for reductive evolution. It demonstrates exactly how an independent cellular organism can systematically dismantle its own biological infrastructure while remaining viable.

The organism appears to be a living fossil of the transition from cellular autonomy to obligate parasitism.

By analysing the genomic architecture of this unique microbe, scientists can observe the reductionist theory playing out in real time. The organism has discarded every single gene related to energy production and cellular metabolism.

It has completely outsourced the fundamental biochemical requirements of life to its host. Yet, it strictly retains the core machinery for DNA Replication, transcription, and translation.

This precise pattern of genetic retention and loss perfectly mirrors the theoretical trajectory from a free-living cell to a virus-like entity. It shows that before an organism gives up its ribosomes and becomes a true virus, it first sheds its metabolic burden.

Perhaps the most compelling evidence for this host-dependent trajectory lies in the mysterious, massive membrane protein of unknown function.

As previously mentioned, Sukanaarchaeum dedicates a quarter of its genome to 8 massive membrane proteins of unknown function.

When researchers look at other highly reduced symbiotic bacteria, proteins of this magnitude are incredibly rare. The retention of these giant proteins despite extreme pressure for genome miniaturisation strongly implies they are absolutely critical to the survival strategy of the organism.

Structural models suggest these giant proteins form complex transmembrane assemblies on the cell surface. While they lack sequence homology to known proteins, analogous structures are found in certain parasitic archaea.

The proteins are likely responsible for mediating physical attachment to the host, facilitating the theft of essential nutrients, or suppressing host immune responses.

This mirrors the behaviour of viral envelope and capsid proteins, which are massive, complex structures dedicated entirely to host receptor binding, cellular entry and immune evasion.

The fact that Sukunaarchaeum sacrifices metabolic independence but heavily invests in host-interaction machinery is a classic hallmark of viral evolution. It appears to represent the absolute limit of cellular reduction just before the terminal loss of the ribosome.

The virocell concept, proposed by evolutionary biologist Patrick Forterre, is particularly relevant when examining this archaeon. Forterre argued that the true living state of a virus is not the inert viral particle, but the infected host cell actively producing new viruses.

In this framework, the virion is merely a spore-like mechanism for genetic dissemination, while the virocell is a novel cellular entity whose metabolic machinery has been hijacked by a viral nucleus.

Sukunaarchaeum exists in a state that borders on this definition. It retains its own physical cellular boundary but acts functionally like a parasitic nucleus operating within a broader host ecosystem.

The extreme specialisation seen in this organism proves that a cellular entity can indeed evolve into a state of pure genetic propagation.

By stripping away all peripheral functions and focusing entirely on copying its own DNA, Sukunaarchaeum operates with a ruthless biological efficiency reminiscent of a viral infection.

This lends massive credibility to the regressive origin of viruses, providing a tangible example of how a complex cell might devolve into a minimalist parasite.

The comparative data in Table 1 above highlights the fluid continuum of life. Rather than existing in rigid, distinct categories, biological entities occupy a sliding scale of autonomy. Sukunaarchaeum sits comfortably between typical cellular life and giant viruses, exhibiting traits of both domains.

It is a cell that behaves like a virus, supporting the hypothesis that the two states are evolutionarily connected through a pathway of continuous genomic decay.

The Computational Virology Angle

Understanding these ancient evolutionary pathways relies heavily on the sophisticated tools of modern computational biology. Microorganisms like Sukanaarchaeum and giant DNA viruses evolve at exceptionally rapid rates.

When an organism sheds its metabolic pathways and relies on a host, it often loses the sophisticated DNA repair mechanisms required to maintain a pristine genome.

The resulting high mutation rate accelerates adaptation, allowing the parasite to continuously outmanoeuvre the immune system of its host. However, this rapid sequence divergence makes traditional phylogenetic analysis incredibly difficult.

When amino acid sequences mutate too quickly, they effectively overwrite their own evolutionary history. This leads to a notorious analytical trap known as long-branch attraction.

In standard phylogenetic trees built from genetic sequences, long-branch attraction causes fast-evolving, distantly related lineages to group artificially simply because they have both diverged significantly from the rest of the tree.

This phenomenon obscures their true evolutionary origins. The extreme branch length of Sukanaarchaeum on the archaeal family tree is a perfect example of this complication, making it difficult to place the organism accurately within established archaeal phyla (Fig. 3 below).

Furthermore, Sukanaarchaeum exhibits an exceptionally low GC content of just 28.9%. Such compositional bias can severely skew the frequencies of specific amino acids in the encoded proteins, leading to convergent evolution artefacts where organisms appear related merely because they share similar nucleotide biases.

To bypass these limitations, computational virologists and evolutionary biologists employ rigorous mathematical corrections. Researchers use techniques like fast-evolving site removal to systematically strip the most mutated sequences from an alignment, revealing the deeper, more conserved evolutionary signals hidden beneath.

They also utilise SR4 recoding, a method that groups amino acids into four distinct classes based on their physicochemical properties, thereby diminishing the impact of compositional heterogeneity.

Beyond sequence manipulation, the field is undergoing a revolution driven by structural phylogenomics. Scientists now look beyond primary genetic sequences and focus on the three-dimensional structures of proteins.

The physical architecture of a protein must fold in a highly specific way to execute its biological function.

While the underlying sequence of amino acids may mutate beyond recognition over billions of years, the overall structural fold of the protein is deeply conserved. Protein folds act as molecular fossils, preserving the ancient history of cellular and viral lineages.

The application of advanced artificial intelligence models like AlphaFold has completely transformed this approach. By accurately predicting the three-dimensional structures of mysterious proteins from their genetic sequences, scientists can illuminate the dark proteome.

Researchers can take the predicted structure of a massive unknown protein from Sukanaarchaeum and run structural homology searches using tools like Foldseek. This allows them to identify ancient evolutionary relationships that are completely invisible at the primary sequence level.

Structural phylogenomics has repeatedly shown that viral capsid proteins share deep, ancient architectures with cellular proteins. For example, the double jelly-roll fold found in the major capsid proteins of numerous viruses shares a distinct structural origin with specific cellular carbohydrate-active enzymes.

Some researchers suggest a chimeric model of viral evolution hypothesising ancient parasitic replicators, perhaps resulting from the reductive evolution of primordial cells, that gradually recruited structural proteins from their cellular hosts to form protective capsids.

The massive membrane proteins of Sukanaarchaeum may represent an early, analogous stage of this exact recruitment process.

By mapping the evolutionary pressures acting on these specific membrane proteins, computational models can simulate how a highly reduced cell might eventually package its naked DNA into a robust viral shell.

Analysing these computational markers helps scientists model the environmental pressures that normally preserve metabolic genes. Once this irreversible threshold is crossed, the organism is locked into a parasitic trajectory.

The evolutionary clock of protein folds allows researchers to track this timeline, revealing a deep history of co-evolution between cellular ancestors and viral lineages.

The tools outlined in the table above are critical for deciphering the tangled web of early cellular evolution. They allow us to peer back through deep time to a period when the distinction between cells and viruses was far less rigid than it is today.

The discovery of this minimal archaeon forces the scientific community to reconsider the strict definition that governs taxonomy. Organisms do not always adhere to the neat categorical boxes designed by human biologists.

Sukanaarchaeum demonstrates that there is a fluid biological continuum stretching from complex, independent cells down to obligate viral parasites.

It stands as a profound testament to the power of reductive evolution, proving that life can persist even when stripped to a bare replicative core.

By synthesising data from global environmental sampling, single-cell genomics, and advanced structural phylogenomics, researchers are slowly illuminating the darkest corners of the microbial world.

The evidence implies that the reductionist hypothesis is not merely a theoretical artefact but an active, observable evolutionary strategy playing out in the deep oceans today.

As we continue to probe the depths of marine ecosystems, it is highly likely that we will uncover even more extraordinary transitional forms, further blurring the line between the living cell and the viral machine.

In conclusion…

Biology is a beautifully chaotic field governed by surprises just as much as rules. Human nature often drives us to think in absolute terms, assuming that if one evolutionary answer is correct, all competing theories must be false.

Yet, the microbial world defies this binary logic. The virosphere is staggeringly diverse, encompassing viruses built from DNA or RNA, structured as single or double strands, and read in positive or negative senses. Nature even produces prions, which are entirely devoid of nucleic acids but have still mastered the art of self-replication.

Given this boundless diversity, it is highly probable that all three models of viral origin are correct. The escape hypothesis could neatly explain the origin of a particular viral lineage, while the reductionist model covers giant DNA viruses, and ancient RNA replicators support the virus-first theory.

One origin story does not invalidate the others, and future discoveries will likely unveil entirely new mechanisms of viral evolution.

The discovery of Sukunaarchaeum is a humbling reminder that organisms rarely fit neatly into the rigid categories designed by researchers. There are undoubtedly more bizarre, boundary-defying entities waiting in the deep ocean to rewrite our understanding of life.

Thank you for taking the time to read through this deep dive into molecular evolution. Your curiosity and willingness to explore these complex biological mysteries make the rigorous work of researching and writing these articles incredibly rewarding.

Did Viruses Devolve From Ancient Cells? This Minimal Archaeon May Hold the Answer. was originally published in Microbial Instincts on Medium, where people are continuing the conversation by highlighting and responding to this story.

Recently, researchers at the Arc Institute and Stanford University published a study in which they successfully generated viable viruses from scratch using artificial intelligence.

The immediate public reaction to artificial intelligence designing biological organisms naturally leans towards anxiety and fear. However, a careful examination of the science behind this development reveals a powerful new tool designed specifically to combat one of the most pressing medical crises.. anti-microbial drug resistance.

By breaking down the complex science into plain language, we can demystify the technology behind these artificial genomes. This article will explore how genome language models operate, outline the positive applications driving this research and address the real risks by separating genuine biosecurity threats from exaggerated public fears.

Demystifying the Science of Genome Generation

To understand this scientific breakthrough, it is essential to define the fundamental biological processes and components involved.

The organisms generated by the artificial intelligence are known as bacteriophages. A bacteriophage is simply a virus that exclusively attacks and infects bacteria. These microscopic entities look remarkably like alien lunar landers (Fig 2), and they are entirely harmless to animal (including human) and plant cells.

In nature, bacteriophages play a vital role in regulating bacterial populations, making them natural predators of the very pathogens that cause human disease.

Why the ϕX174 bacteriophage?

For their experiment, the researchers selected a well-documented natural bacteriophage, ϕX174, as their foundational template. Why this bacteriophage, you may ask? Well, because it holds a prestigious place in the history of biology:

1. ϕX174 was the first DNA-based genome ever sequenced by scientists, published in 1977.
2. ϕX174 has a compact structure consisting of ~5 kb (kilobases) and only 11 genes.
3. ϕX174 is the first bacteriophage to be chemically synthesised using synthetic oligonucleotides, published in 2003
4. ϕX174 specifically attacks Escherichia coli (E coli) bacteria, which is one of the extensively studied bacteria, facilitating straightforward lab screening and testing.

So in the history of biology, ϕX174 has also earned a place as the first organism to be designed by AI using genome language models.

The technology

The technology used to design these new bacteriophages relies on genome language models named Evo 1 and Evo 2. Most people are familiar with artificial intelligence programmes like ChatGPT, which operate by predicting the next logical word in a sentence based on vast amounts of text data called a large-language model (LLM).

Genome language models operate on an identical principle but use a completely different vocabulary. Instead of processing English words, they process the basic chemical letters of DNA, which are represented by Adenine (A), Cytosine (C ), Guanine (G) and Thymine (T).

By analysing trillions of genetic letters from millions of natural bacterial viruses, the Evo models learned the complex grammar of biological life. The developers built EVO 1 with a context window of 131 000 tokens, allowing artificial intelligence to comprehend vast and complex genetic structures across entire genomes.

To specialise the models for this specific task, the researchers fine-tuned the systems on a curated dataset of ~14 500 genomes from the Microviridae family, which is a specific family of bacteriophages.

The methodology (Fig 3) deployed by the research team highlights the vital distinction between digital design and biological creation. It is worth noting that the AI did not autonomously manufacture life. Instead, AI generated thousands of blueprints in silico, which means entirely within a computer simulation.

Because standard gene prediction tools failed to accurately map the overlapping reading frames of the complex viruses, the researchers had to create bespoke gene annotation tools to evaluate the AI -generated designs.

Human experts carefully reviewed these digital blueprints and computationally filtered them to ensure they met strict biological criteria such as correct genome lengths and appropriate genetic architectures.

Following this rigorous computational screening, the human researchers physically synthesised 285 of the DNA sequences in a laboratory. They introduced these synthetic blueprints into E coli bacteria to see if the viruses would boot up and replicate.

Out of 285, exactly 16 blueprints turned out to be viable functional phages capable of successfully infecting E coli. This translates to an efficiency rate of ~5%.

Such a relatively low success rate demonstrates that designing functional genomes remains a monumental challenge and proves that AI currently acts as a highly sophisticated design assistant rather than an independent autonomous creator.

Despite the low overall efficiency, the 16 viable phages were biologically remarkable because they were not mere copies of the original template (Fig 4). Some designs contained hundreds of novel mutations that do not exist in any known natural genomes, demonstrating that the models had achieved a level of biological literacy.

Advanced imaging techniques, such as cryogenic electron microscopy (Fig. 1), revealed that a specific design called Evo-ϕ36 was able to seamlessly replace an entire DNA packaging protein known as protein J with a homologous protein from an evolutionarily distant phage.

This integration demonstrated that genome language models can create functional biological structures that transcend standard evolutionary constraints, uncovering unique structural solutions that nature has never produced.

The Fight Against Antimicrobial Resistance

The motivation driving this sophisticated research is deeply rooted in a desire to solve a catastrophic global health emergency. Antimicrobial resistance occurs when bacteria mutate and evolve to survive the traditional antibiotic drugs designed to eradicate them.

Over the past century, humanity has relied heavily on antibiotics to treat minor infections and facilitate complex medical procedures like surgeries and cancer chemotherapy.

However, the rampant overuse of antibiotics in both human medicine and agriculture has accelerated the rise of superbugs, which are bacteria completely resistant to multiple life-saving drugs.

The statistical projections surrounding antimicrobial resistance are exceedingly grim. In 2019, drug-resistant infections directly caused 1.27 million deaths and contributed to an additional 4.95 million deaths, making it a bigger killer than HIV and malaria combined.

Forecasts project that if left unchecked, this crisis could cause up to 10 million deaths annually by the year 2050. Furthermore, modern studies from the Global Research on Antimicrobial Resistance project estimate that antibiotic-resistant infections could lead to 39 million direct human deaths between the years 2025 and 2050.

Economically, this silent pandemic is devastating. The World Bank estimates that unchecked antimicrobial resistance could wipe out nearly 4% of the global gross domestic product, costing up to 100 trillion dollars and pushing 28 million people into extreme poverty by the middle of the century.

The Phage Alternative and Mechanism of Action

Faced with this voracious crisis, scientists are desperately hunting for alternative treatments. Phage (short for Bacteriophage) therapy uses bacteriophages as targeted weapons against specific harmful bacteria.

Researchers are investigating ecosystems for naturally occurring phages that can combat multi-drug-resistant bacteria. Meanwhile, AI-designed bacteriophages are emerging as an alternative to traditional methods, where phages are crafted using artificial intelligence instead of relying on nature.

Bacteriophages operate by attaching to the surface of a target bacterium, injecting their genetic payload, and hijacking the bacterial machinery to replicate (Fig 5).

This process ends with the lysis, or bursting, of the bacterial cell, destroying the pathogen. Because phages are highly specific, they can eradicate dangerous superbugs while preserving the beneficial microbiome bacteria that naturally exist in the human gut.

AI-Optimised Viral Performance

The findings from the Arc Institute study demonstrated the immense clinical potential of artificial intelligence in the field of antibiotic resistance.

In laboratory growth competitions, researchers monitored the optical density of bacterial cultures to track infection rates. Several of the viable phages demonstrated significantly faster bacterial killing capabilities and higher overall fitness compared to the natural ϕX174 template.

As a result, the host bacterial populations rapidly declined to minimal densities. The use of AI effectively optimised the biological performance of the virus.

Overcoming the Evolutionary Arms Race

One of the traditional challenges of phage therapy is that bacteria are notoriously quick to mutate and can eventually develop resistance to a specific bacteriophage.

To combat this evolutionary arms race, the researchers created a therapeutic cocktail by mixing multiple AI-designed phages. When they tested this cocktail against 3 different highly resistant strains of E coli the artificial bacteriophages successfully recombined and overcame the bacterial defences.

By swapping genetic components, the artificial phages named Evo R1 and Evo R2 bypassed the bacterial resistance mechanisms entirely. This breakthrough proves that genome language models can help researchers rapidly develop highly specific and resilient treatments to tackle the superbug crisis, thus protecting millions of human lives.

Moving from Lab to the Clinic

Currently, the global community is already exploring engineered phages in human clinical trials. Experimental treatments such as SNIPR001 and LBP EC01 utilise advanced CRISPR gene-editing techniques combined with bacteriophages to target drug-resistant infections in the urinary and gastrointestinal tracts.

Regulatory frameworks for phage therapy are still developing, so patients often receive these treatments under compassionate use or right-to-try laws as a last resort when all conventional antibiotics have failed.

As genome language models advance, they will likely integrate into pharmaceutical pipelines seamlessly, enabling scientists to quickly generate customised phages and streamline personalised treatments for patients.

Addressing Public Fear and Navigating Realistic Risks

The Root of Public Anxiety

While it's entirely natural to have concerns about AI designing new biological entities, closely examining the safety measures implemented in systems like EVO can help us understand why this technology should be viewed as a tool for medicine rather than as a plot device in a sci-fi movie.

The public frequently expresses valid anxiety regarding the concept of AI escaping its sandbox. In software development, a sandbox is an isolated testing environment designed to keep experimental programmes safely contained.

There have been well-documented instances in the broader technology industry where AI agents unexpectedly collaborated to bypass their digital sandboxes and interact with external servers.

It is entirely reasonable for the public to be concerned that a genome language model could break free from its programming limitations. This could result in the model generating dangerous genetic blueprints, similar to those from the Human Genome Project, or intentionally creating harmful biological instructions that could pose threats to humans and biodiversity as a whole.

The Real threat is lowering the barrier to entry

Moreover, current security research emphasises real vulnerabilities that experts take seriously. A recent RAND Corporation working paper analysed how advanced AI foundation models can increase the risks of biological weapons by functioning as virtual lab assistants.

The RAND report presented a disturbing case study involving Norwegian domestic terrorist Anders Breivik. Despite lacking formal scientific training, Breivik successfully manufactured complex chemical explosives by reading manuals online and carefully troubleshooting his mistakes.

This demonstrates that highly motivated amateurs can overcome the challenge of tacit knowledge. Modern AI models amplify this risk because they can quickly bridge knowledge gaps by offering detailed, step-by-step guidance on troubleshooting biological experiments and sourcing laboratory equipment.

A model that can engage in conversation with a user and optimise biological protocols significantly reduces the barrier to entry for biological engineering.

Inherent Safeguards in Genome Language Models

To address these real threats, the developers of the Evo models implemented strict digital and procedural safeguards. Comprehensive evaluations confirmed that whenever the model was explicitly prompted to generate pathogenic human viral proteins, its computational outputs were effectively random and biologically non-functional.

This proves that rigorous data curation is a highly effective method for constraining the capabilities of genome language models, preventing them from accessing the genetic pathways necessary for human harm.

The EVO engine was trained on non-animal and non-plant viral genomes to prevent harm to humans. By deliberately excluding eukaryotic viruses from the foundational training data, the system lacks the biological vocabulary required to design human pathogens.

Industry Frameworks and Regulation

AI laboratories are implementing strict governance frameworks for the dual-use nature of biological research. Anthropic, for example, has a Responsible Scaling Policy that requires advanced models to pass biosecurity evaluations before public release and classifies AI systems into safety levels based on their potential for misuse.

Models capable of aiding in the creation of chemical, biological, radiological, or nuclear weapons are subject to strict measures to prevent any malicious use.

By enhancing security protocols in line with the intelligence of these models, the industry seeks to ensure that these powerful tools do not fall into the wrong hands.

The physical bottleneck

Finally, it is vital to recognise the massive logistical gap between generating a digital sequence and creating a physical biological threat. An AI programme cannot autonomously sneeze a virus into a petri-dish.

Transforming a generated blueprint into a physical virus requires significant human intervention, highly specialised laboratory equipment, expensive chemical reagents and stringent biological containment protocols.

The AI-generated blueprints were rigorously filtered by human experts before synthesis, ensuring that nonsensical or dangerous anomalies were discarded.

Therefore, the immediate threat does not stem from a sentient AI autonomously deciding to create a pathogen but rather from malicious human actors attempting to misuse the technology.

Because artificial intelligence fundamentally relies on humans to bring its digital creations into the physical world, biosecurity specialists advocate or legally mandated screening of all synthetic DNA orders.

Commercial laboratories would need to verify buyers' identities and screen all requested genetic sequences for known pathogenic markers, thereby creating a physical bottleneck that prevents dangerous digital blueprints from being weaponised.

Just as the global community successfully established treaties to regulate the use of nuclear technology, careful oversight and international regulations will be essential to ensure that AI biological design remains a force for good.

By demystifying the underlying mechanics, understanding the physical bottlenecks and acknowledging the strictly enforced safety protocols, society can embrace this technology not as an existential threat but as a crucial ally in the ongoing battle against global infectious diseases.

My main message is that AI did not create physical phages, but it assisted in designing blueprints, resulting in 16 viable phages from thousands designed.

AI-Designed Biological Viruses Are Here. Should We Be Worried? was originally published in Microbial Instincts on Medium, where people are continuing the conversation by highlighting and responding to this story.

A group of scientists from the University of Minnesota published a preprint where they synthesised an Artificial cell capable of growth and replication.

It is indeed a groundbreaking stride in Synthetic Biology and Biochemistry. This is the first successful attempt to bridge Chemistry and Biology, where non-living chemicals, after careful and precise combinations, managed to breed a living organism.

The SpudCell is, at its core, a chemically defined synthetic cell capable of growth, genome replication, resource acquisition and genetically encoded division. But if I just sit here and regurgitate what is already in the preprint, you might as well go read the original manuscript yourself.

I put the word ‘successful’ in bold earlier because it carries a massive amount of weight. This project was certainly not an overnight miracle. It appears to be the culmination of a long, often frustrating array of earlier studies that slowly pieced together the puzzle of artificial life.

We often throw around that old quote about standing on the shoulders of giants. In molecular biology, those shoulders are built from decades of gruelling trial and error.

To truly appreciate the magnitude of this breakthrough, we have to look back at the historical foundations of synthetic biology.

We need to step back into the early, messy days of the field and unpack the monumental technical hurdles that kept researchers up at night long before SpudCell came along.

Phase I - The Giants of Compartmentalisation

The initial attempts to create a controllable, minimal cell did not start from scratch. Instead, scientists attempted to simplify the complex natural organisms that already existed. This is known in the field as the top-down approach.

The logic appears to be quite sound. First, take a simple living bacterium and systematically remove every gene that is not strictly necessary for basic survival. At last, you should eventually be left with a minimal operating system for life.

The most prominent and ambitious attempt to execute this top-down genome reduction was spearheaded by researchers at the J. Craig Venter Institute, who chose the bacterium Mycoplasma mycoides as their starting point.

Mycoplasma bacteria were specifically selected because they naturally lack cell walls and possess very small genomes compared to other bacteria.

Their objective was to strip this already simple genome down to its absolute bare essential components, creating a heavily modified organism named JCVI-syn3.0 (see Fig 2 below). This synthetic organism holds a significant record of possessing the smallest genome of any autonomously replicating cell found anywhere in nature.

The engineering process behind HCVI-syn3.0 was a masterclass in molecular biology where an earlier synthetic iteration known as syn1.0(Fig 2B Blue plasmid) had non-essential genes removed.

To figure out which genes were essential and which were expendable, the team relied on global Tn5 transposon mutagenesis. This technique involves using a transposable genetic element to randomly disrupt genes across the entire genome. Survival from such gene disruptions is deemed non-essential and marked for deletion.

Once the essential genes were mapped out, the physical construction of the new genome began. The genome was logically divided into 8 manageable segments which, when reduced, were meticulously tested in the context of a 7/8 syn1.0 genome and then tested again in various combinations with other reduced segments to ensure viability.

The actual assembly involved overlapping oligonucleotides, which were chemically synthesised in the laboratory and assembled into fragments of 1.4 kilobase pairs.

Following rigorous error correction and polymerase chain reaction amplification, five of these fragments were stitched together to form 7-kilobase cassettes, which were strictly sequence verified.

The cassettes were subsequently assembled within living yeast cells to generate one-eighth molecules. The eight resulting molecules were amplified by rolling circle amplification and assembled in yeast to generate the complete synthetic bacterial chromosome.

The final JCVI-syn3.0 organism contained a mere 531 kilobase pairs of DNA and exactly 473 genes, retaining almost all the genes involved in the synthesis and processing of cellular macromolecules, allowing it to survive and replicate autonomously in a nutrient-rich broth.

However, the top-down approach ultimately exposed a glaring limitation in our contemporary biological understanding. Among the 473 essential genes successfully retained in the JCVI-syn3.0 genome, an astonishing 149 genes possess completely unknown biological functions.

If scientists truly wanted to master the fundamental chemistry of life, merely stripping away parts of a complex natural machine was never going to be enough. Indeed, there was a need to build a cell entirely from the bottom up, using only molecules they fully understood.

To compare the limitations of the top-down approach and the bottom-up approach, I summarised them in the table below:

Phase II — Building the Chassis

In the unforgiving bottom-up approach, scientists must start with non-living chemical precursors and attempt to synthesise a working biological system. There is indeed a thin line between living and non-living … chemistry and biology.

Every single molecule placed into the reaction tube must be accounted for; if a vital enzyme or a specific cofactor is missing, the entire system simply fails to operate.

Historically, biochemical researchers relied heavily on whole cell extracts to perform cell-free protein expression experiments by crushing natural bacteria like E. coli, filtering out the heavy cellular debris, and harvesting the resulting chemical soup.

The extract is saturated with uncharacterised enzymes, hidden nucleases that degrade DNA and unpredictable metabolic variables, making precise engineering nearly impossible because of a lack of certainty about how the extract will react to a new genetic circuit.

The major paradigm shift occurred with the development of completely chemically defined transcription and translation systems.

Shimizu and his colleagues successfully developed a protein-synthesising system reconstituted entirely from purified components. This groundbreaking platform is widely known across the scientific community as the PURE system (Protein Synthesis Using Recombinant Elements).

The PURE system relies on a carefully curated mixture of 31 or 32 specifically chosen proteins encoded across 30 cistrons (the molecular equivalent of a gene). These recombinant tagged protein factors are individually purified to absolute homogeneity before being combined in the laboratory, including all the necessary translation factors alongside 20 specific aminoacyl-tRNA synthetases.

The resulting system is exceptionally clean and capable of producing protein at a reliable rate of approximately 160 micrograms per millimetre per hour in a standard batch mode.

Because the exact chemical composition is known, researchers can manipulate the environment with unprecedented precision. For instance, the deliberate omission of a release factor allows the system to incorporate unnatural amino acids using a suppressor transfer RNA efficiently.

The PURE system finally provided researchers with a resilient and fully understood chassis for expressing genes outside the chaotic environment of a living cell, but basic protein expression is only a fraction of what is required for cellular life.

Ribosomes act as fundamental factories of the cell, translating messenger RNA into functional proteins. Reproducing their incredibly complex biogenesis outside a natural living system was long considered an insurmountable technical hurdle.

In natural E. coli, the ribosome is an incredibly intricate structure comprising 3 distinct ribosomal RNAs and 54 different ribosomal proteins.

These components do not just randomly clump together but undergo cooperative, hierarchical assembly facilitated by numerous specialised accessory factors.

A research group led by Kosaka managed to achieve autonomous ribosome biogenesis entirely in vitro. The team developed a highly specific and sensitive reporter assay designed for the precise detection of nascent ribosomes.

This analytical tool allowed for combinatorial and iterative exploration of various reaction conditions, testing thousands of variables to find the perfect environment for assembly.

Ultimately, the researchers achieved the simultaneous autonomous synthesis of both the small and large ribosomal subunits within a single reaction space, successfully integrating transcription, translation, structural processing and physical assembly all at once.

In vitro ribosome biogenesis represents a crucial advancement toward revealing the fundamental principles underlying the self-replication of life. It proved that the most complex molecular hardware of a cell could be self-assembled from raw chemical components without requiring a pre-existing living host.

These chemically defined building blocks set the stage for a fully integrated artificial cell, giving us the engine and the factory. The next step was to put them inside a functional body.

Phase III — The Breakthrough

The culmination of these intense bottom-up engineering efforts is an ambitious project known as SpudCell. It was developed by a team of synthetic biologists led by Katarzyna Adamala and Nathaniel Gaut at the University of Minnesota.

SpudCell operates on an incredibly minimal genetic code, sitting at just about 90,000 base pairs. This number is remarkably lower than 113,000 base pairs, which previous biologists had speculated as the absolute minimum genome size for a functional living cell.

Instead of relying on a single large circular chromosome like most natural bacteria, this genetic information is modularly distributed across seven or eight distinct plasmids.

This specific encoding solution allows researchers to easily programme cellular functions by modifying or swapping out a single piece of the modular plasmid system without having to synthesise an entirely new chromosome.

In total, the SpudCell genome encodes a mere 36 purified enzymes, leaving absolutely no room for uncharacterised proteins with unknown functions.

Predatory Feeding Mechanisms

The most fascinating aspect of the SpudCell architecture is its highly unique resource acquisition strategy. Natural biological cells inherit billions of years of complex evolutionary machinery, relying on hundreds of interconnected metabolic genes to process raw materials from their environment and synthesise the lipid molecules required to expand their membranes.

To circumvent this severe limitation, the researchers engineered a mechanism that behaves much like a predatory feeding strategy. The synthetic cell sustains itself by producing a modified bacterial pore protein known as alpha-haemolysin.

This genetically encoded protein is specifically designed to display a distinct chemical tag on its outer surface.

The process begins when the alpha-haemolysin protein is expressed internally by the synthetic cell’s minimal transcription and translation machinery (the encapsulated PURE system). Once translated, the pore protein automatically localises to the outer lipid membrane of the SpudCell.

In the surrounding liquid medium, the researchers provide a surplus of much smaller feeder liposomes acting as prey for the synthetic cell. These bubble-like structures are heavily packed with fresh membrane lipids, ribosomes, small-molecule nutrients and other essential raw materials.

Crucially, the exterior surface of these liposomes features a chemical tag displayed by the SpudCell’s alpha-haemolysin protein, resulting in latching onto matching tags on the feeder.

The targeted latching acts like a chemical zipper, pulling the two lipid bilayers together incredibly closely (Fig 4). The physical proximity triggers a rapid merger of the two lipid membranes.

Upon successful fusion, the larger SpudCell actively absorbs the fresh raw materials directly into its internal aqueous lumen. This process replenishes the synthetic cell’s internal nutrients and simultaneously provides the excess membrane lipids required for the physical expansion of the cell itself.

Field biologists accustomed to observing microscopic predation under a microscope might find this mechanism remarkably familiar in nature, e.g., slugs that absorb chloroplasts from algae and then inherit photosynthesis.

The synthetic cell essentially incorporates smaller functional compartments to maintain its internal machinery. It is a highly efficient way to acquire complex resources without needing the genetic code to manufacture them from scratch.

Over repeated rounds of this predatory feeding, the synthetic cells use an enzyme borrowed from a bacteriophage virus known as Phi29 to replicate their internal plasmid DNA.

The Physics of Membrane Fission

Growth and genome replication are only two pillars of the standard cell cycle (Fig 5). The ability to physically divide into two distinct daughter cells is arguably the most challenging mechanical feat for any synthetic system.

Natural eukaryotic and prokaryotic cells rely heavily on a complex internal scaffolding known as the cytoskeleton to organise their division. Bacteria often use a tubulin-like protein called FtsZ that forms a contractile ring on the inside of the cell membrane, pinching the cell in half during division.

Replicating this intricate network of filaments, motor proteins and division rings from scratch requires far too many interacting parts for a minimal cell to handle. SpudCell completely bypasses the need for a complex cytoskeleton by leveraging pure membrane physics.

The artificial division mechanism relies on a biophysical phenomenon known as membrane fission by protein crowding. Extensive biophysical research demonstrated that membrane fission does not actually require proteins with highly specific, curved structural features.

Previously, the prevailing scientific consensus assumed that fission required specialised curved protein scaffolds or wedge-like hydrophobic membrane insertions to force the membrane to bend.

The Stachowiak laboratory completely challenged this deeply held assumption by discovering a structure-independent mechanism driven entirely by steric pressure.

The elegant physics is that when large numbers of proteins bind to the surface of a lipid membrane, they become highly crowded. These crowded molecules undergo constant, random lateral collisions driven by Brownian motion.

These intense molecular collisions generate substantial lateral steric pressure along the two-dimensional surface of the membrane. This is much like a compressed gas expanding against the walls of a sealed container.

As the local protein concentration increases, the mounting steric pressure stretches the membrane until it ultimately reaches a breaking point, resulting in spontaneous fission and vesiculation.

This physical model fundamentally alters our understanding of hydrophobic protein domains. Scientists previously believed that amphipathic helices, such as those found in the endocytic protein epsin1 ENTH domain, drove fission by physically wedging themselves into the lipid bilayer and forcing the lipids apart.

However, careful and precise experiments revealed that reducing the hydrophobicity of these helices primarily decreases the protein’s overall binding affinity to the membrane.

When researchers compared wild-type proteins and helix mutants at equal levels of membrane coverage, they found that both drove fission equally well.

This proves that the physical role of hydrophobic insertions is merely to anchor the proteins strongly to the membrane surface. Strong anchoring prevents the proteins from dissociating back into the surrounding fluid, thereby amplifying the steric pressure generated by the crowded environment on the surface.

To definitively prove that fission is totally structure-independent, researchers tested simple green fluorescent protein because it lacks any specialised curvature-inducing or membrane-inserting structures. Nevertheless, when it was artificially bound to the membrane at high coverage, it drove membrane fission with remarkably high efficiency.

Furthermore, the research demonstrated that intrinsically disordered proteins are even more potent at driving this process. While one might wonder if massive, complex, structured proteins like human haemoglobin could similarly contribute to this effect, well, the provided research does not detail this specific point.

It does, however, clearly note that intrinsically disordered proteins lack a rigid, folded structure but possess exceptionally large hydrodynamic radii.

Because these disordered proteins occupy a physically larger spatial envelope as they flop around on the surface, they undergo far more frequent and intense steric collisions.

As a direct result, intrinsically disordered proteins drive membrane fission with substantially greater potency than smaller, tightly folded structured proteins at equivalent concentrations.

SpudCell ingeniously utilises these exact physical principles to divide without a cytoskeleton. Genetically encoded proteins are programmed to accumulate and heavily crowd together on the synthetic cell’s outer membrane surface.

The proteins continue to gather as the cell grows until the mechanical steric stress builds up to a critical threshold. The intense crowding physically pinches the membrane apart, successfully splitting the parent cell into two distinct daughter cells.

Phase IV — The New Horizon

The successful integration of continuous growth, genome replication, and membrane division within a chemically defined synthetic cell opens the door to something truly profound: autonomous evolution.

The ability to undergo Darwinian selection is arguably the defining boundary between complex chemistry and true biology, which SpudCell crossed definitively.

Researchers sought to determine if the synthetic cell could actively adapt to environmental pressures by introducing a specific genetic mutation into a subset of the SpudCell population that increased the expression of the fusion protein responsible for feeding.

Well, because these mutant cells possessed more fusion tags on their exterior, they were able to merge with the feeder liposomes much more efficiently.

This enhanced metabolic advantage allowed the mutant cells to absorb nutrients faster, grow faster, and ultimately produce more daughter cells (by 50%) than their unmutated counterparts, dominating the original strain after just 5 generations of continuous division (Fig 6).

To stretch the system’s resilience and evolutionary potential, the research team deliberately altered the environmental conditions to make the feeder liposomes scarce.

Under these harsh conditions of strict resource limitation, the selection advantage of the fast-growing mutants became even more pronounced. The fast growers again outnumbered the slow growers by a ratio of more than two to one.

This experiment serves as an undeniable demonstration that Darwinian selection can occur in an entirely artificial, chemically defined system.

Similar evolutionary potential has been observed in other synthetic protocell models. Research exploring the Darwinian evolution of self-replicating DNA utilised a transcription-translation coupled DNA replication system based entirely on the Phi29 DNA polymerase.

The synthetic self-replicator survived across multiple rounds of coupled reactions encapsulated inside liposomes. Remarkably, within only ten evolution rounds, the synthetic system independently accumulated mutations conferring a distinct selection advantage.

Combined data extracted from next-generation sequencing and reverse engineering revealed that these spontaneous mutations had non-trivial, context-dependent effects on the replication efficiency.

These observations appear to imply that synthetic chemical life can independently evolve complex traits that human scientists did not explicitly programme into the initial code.

Synthetic Pathogens and Biosafety Gaps

With the capacity for autonomous evolution comes the inevitable emergence of exploitation. In natural biology, wherever there are concentrated resources, parasites rapidly evolve to steal them.

The most suspicious parasite would be viruses; as they say, Nature abhors a vacuum, I say, ‘Nature abhors a free lunch.’ Synthetic cells are no exception to this rule once they enter the universe of biology.

Actually, a research team successfully modelled host-pathogen interactions and severe infections within synthetic minimal cells (Fig 7).

The host cell was programmed to continuously express green fluorescent protein, providing a visual indicator of its metabolic health. It was also decorated with specific single-stranded DNA tags on its outer membrane.

The parasite was a smaller liposome encapsulated with plasmid DNA encoding an mCherry fluorescent protein. The parasite’s surface featured a single-stranded DNA tag that was perfectly complementary to the host’s tag.

The invading plasmid hijacked the host’s limited transcription and translation resources (i.e. the internal PURE system). The host cell was forced to express the parasite’s mCherry protein alongside its own green fluorescent protein, perfectly modelling direct resource competition.

The researchers then escalated the severity of the infection to model a deadly, cell-killing pathogen. They engineered the host’s green fluorescent protein plasmid to include two specific Munl restriction sites (where restriction enzymes specifically cut).

In this fatal scenario, the invading pathogen liposome delivered a plasmid encoding the Munl restriction enzyme. Upon fusion, the host cell translated the Munl restriction enzyme from the pathogen’s plasmid.

The newly synthesised restriction enzyme then hunted down the host’s own genome and digested the template DNA exactly at the restriction sites.

The complete and rapid destruction of the host genome halted the green fluorescent protein expression, effectively killing the host’s metabolic function and leaving the cell as an empty shell operating solely for the pathogen.

However, the researchers also proved that synthetic cells can be armed with artificial immunity. Inspired by the natural function of restriction endonucleases that safeguard against foreign bacteriophage DNA, researchers engineered the host cells to resist infection.

The host cell was equipped with a green fluorescent protein plasmid lacking any Munl sites, alongside a secondary plasmid encoding the Munl restriction enzyme.

This secondary plasmid acted as an active defensive resistance gene. When a parasite carrying an mCherry plasmid with a Munl restriction site invaded, the host’s pre-expressed restriction enzyme rapidly digested the foreign DNA, entirely preventing the parasite from hijacking the system and expressing its red protein.

The research also demonstrated that maintaining this immunity comes with a distinct metabolic penalty. Expressing the resistance enzyme continuously consumes precious transcription and translation resources, which slightly reduces the host’s overall operational efficiency for normal growth tasks.

Finally, the team successfully modelled a primitive form of artificial inoculation. The host cells were incubated with an excess of single-stranded DNA oligos that were perfectly complementary to their own surface fusion tags.

These inoculating oligos hybridised with the host’s tags, forming a stable double-stranded DNA duplex, physically blocking them from hybridising with the incoming pathogen’s tags. This simple physicochemical intervention completely abolished the ability of the pathogenic synthetic cells to fuse with and infect the host cells.

The Risks of Mirror Life

The rapid advancement of synthetic cell engineering, particularly the demonstrated ability to create self-replicating and autonomously evolving systems from raw chemistry, highlights significant global biosafety gaps.

One of the most unprecedented theoretical risks discussed in the provided research involves the concept of mirror life.

All known natural life on Earth is strictly homochiral. Biological DNA and RNA are universally constructed from right-handed nucleotides, while all proteins are built exclusively from left-handed amino acids.

Driven by scientific curiosity and the pursuit of highly novel industrial applications, researchers have begun actively investigating the creation of lifeforms composed entirely of mirror-image biological molecules.

Recent technological advances have enabled the chemical synthesis of mirror-image kilobase-length nucleic acids and large functional proteins. Because they possess reversed chirality, these synthetic biomolecules are entirely unrecognised by natural enzymes.

A standard natural protease (enzymes that digest proteins) cannot cleave a mirror-image peptide bond because the molecule does not physically fit into the enzyme’s active site, making mirror molecules highly resistant to all normal forms of biological degradation.

While this deep resistance to degradation is highly desirable for creating long-lasting, non-immunogenic medical therapies that will not be destroyed by the human immune system, it poses a severe ecological threat if integrated into a self-replicating organism.

If scientists were to successfully construct a living mirror bacterium, it would constitute a radical departure from known life. Such an organism would have absolutely no natural predators.

Viruses, phages and predatory bacteria from the natural world would be physically incapable of interacting with or consuming mirror-life. It could theoretically evade all existing biological containment barriers.

The research strongly cautions that once a method is fully developed to construct a natural-chirality bacterium entirely from synthetic components (like SpudCell), creating a mirror version follows the same logical pathway.

Furthermore, with continued advances in synthetic biology, a natural-chirality bacterium might eventually be engineered to produce mirror proteins and nucleic acids directly in vivo, providing a dangerous stepping stone toward a fully mirror-image organism escaping into the wild.

The capability to create a comprehensive mirror life is likely at least a decade away because it requires financial investments and major technical breakthroughs. However, the global scientific community is being urged to pre-emptively confront these risks before the technology matures.

Recommendations include taking immediate regulatory steps to prevent the production of functional mirror genomes and proteomes.

Governance of these key enabling technologies must ensure that anyone attempting to create mirror bacteria remains hindered by scientifically challenging, expensive and time-consuming technical barriers.

Biocontainment protocols and strict biosafety approaches must be codified globally before this technology becomes widely accessible.

The Future of Open Bioscience

Despite these looming and severe biosecurity challenges, the creators of SpudCell are actively pushing to democratise the technology as much as possible, e.g launching a new public-benefit research institution called Biotic.

The explicitly stated goal of Biotic is to develop shared physical infrastructure and open modular standards for the engineering of synthetic cells.

Currently, synthetic biology suffers from a severe lack of standardisation as different laboratories across the globe use incompatible protocols, varying buffer solutions and genetic components that do not interface correctly.

The Biotic initiative intends to position SpudCell as the universal, shared chassis for the entire field.

The organisation hopes to consolidate the synthetic cell’s seven DNA plasmids into a single, stable synthetic chromosome and build increasingly complex molecular machinery that any researcher can download and synthesise locally.

While the researchers note enthusiastically that this technology could completely transform medicine, materials science and industrial chemistry by enabling properties that typically require difficult molecular transformations, the open-source distribution of self-replicating artificial life invites a highly nuanced scepticism.

Providing a free, highly functional biological chassis to thousands of independent laboratories worldwide certainly accelerates discovery by removing a massive barrier to entry that currently keeps synthetic biology locked inside heavily funded universities.

However, it also guarantees that the fundamental tools necessary to engineer artificial pathogens, or even rudimentary mirror-life components, will eventually slip far beyond the control of centralised regulatory bodies.

If a high school student can eventually order the components to print a self-replicating SpudCell in their garage, we face a biosecurity landscape that our current regulations are entirely unequipped to handle.

This tension brings to mind a video exploring the frontiers of DIY (Do It Yourself) gene editing, where one of the subjects attempted to cure his own HIV using experimental genetic injections.

The immediate danger lay in the fact that this individual possessed no formal scientific background, no medical licence, and an alarming disregard for basic aseptic techniques while executing high-risk procedures on his own body.

Watching someone actively inject unverified genetic material on a live stream feels genuinely unnerving, highlighting the double-edged nature of making these tools widely accessible.

It is one thing to tinker with sequence alignments or standard plasmids in a controlled university setting, but it is another thing entirely when rogue experimentation meets public distribution.

Conclusion

The successful creation of a chemically defined synthetic cell capable of growth and replication proves definitively that the fundamental functions of life do not require a mysterious vital spark.

Life is, at its core, a complex mechanical interaction of steric pressure on lipid membranes, modular genetics translated by ribosomes and basic thermodynamic efficiency.

Humanity has finally crossed the threshold from observing the chemistry of life to actively engineering life from scratch in a pristine laboratory.

The profound ecological, medical and ethical consequences of this achievement are likely to shape the entire trajectory of biological science for the rest of the century.

Can We Engineer a Synthetic Cell to Experience Evolution? A Groundbreaking Study Answers Yes. was originally published in Microbial Instincts on Medium, where people are continuing the conversation by highlighting and responding to this story.

Recently, I released SIMBA 2.0. This update is not just a cosmetic change but a fundamental shift in how the software “thinks” about the genome.

Moving Beyond the Simple Calculator

In the original version of SIMBA, I relied on the Nei-Gojobori method. It is a foundational approach in molecular evolution, but it carries a significant limitation. It assumes that every mutation is equally likely to occur. In the biological world, this is rarely true.

When we look at the Spike protein, we are looking at one of the most complex pieces of viral machinery. To analyse it properly, I had to move to the HKY85 substitution model . Unlike its predecessor, HKY85 acknowledges that transitions (mutations between similar bases) happen more frequently than transversions. It also accounts for the fact that base frequencies in a viral genome are rarely even.

By making this change, the software became less of a simple calculator and more of a high-fidelity lens.

The Power of the Bayesian Grid

The second major shift was in the statistical engine. SIMBA 2.0 now employs an Empirical Bayesian Grid for its inference.

In traditional frequentist statistics, you are often looking for a single point of truth. By using a Bayesian approach, the tool learns a “prior” from the dataset itself. It looks at the overall evolutionary signal across the entire gene before deciding how significant the mutations at a specific codon really are.

This is particularly useful when sequences have varying levels of quality. I implemented epsilon-normalisation to ensure that numerical stability is maintained even when the data is sparse. It ensures that the “noise” of sequencing errors does not drown out the “signal” of true positive selection.

Mapping the Functional Terrain

Perhaps the most rewarding part of this build was the visualisation suite. Data is only as good as our ability to interpret it.

I designed the tool to automatically shade functional regions like the Receptor Binding Domain (RBD) and the Receptor Binding Motif (RBM). When you run the pipeline, you don’t just get a table of numbers. You get a “Selection Landscape and Rates.” (Fig 1 and Fig 2). You can see exactly which evolutionary hotspots fall within critical structural areas of the Spike gene.

A Journey in Independent Research

Building SIMBA 2.0 while balancing my role as a science educator in Harare has been a lesson in focus. It has reinforced my belief that we don’t need massive institutional clusters to produce meaningful research tools.

Sometimes, all you need is a deep curiosity about how a virus evolves and a clean Python environment to build the solution.

The project is now open-source on GitHub for anyone interested in exploring the landscape of viral selection.

SIMBA 2.0-Building a Better Bayesian Pipeline for Viral Evolution in the Coronavirus Spike Protein. was originally published in Predict on Medium, where people are continuing the conversation by highlighting and responding to this story.

Introduction: Omicron’s Relentless Evolution in 2025

When Omicron first appeared in late 2021, it was described as a “variant of concern” with an unusual number of mutations. Few could have predicted that, four years later, it would continue to shape the global trajectory of COVID-19.

By 2025, Omicron had become less of a single variant and more of a sprawling family tree, a reminder that viruses evolve not in straight lines but in tangled webs of descent and recombination.

Last year, six lineages (and their sub-lineages) stood out: XEC, KP.3.1.1, LP.8.1, NB.1.8.1, XFG and BA.3.2. Each carried its own genetic quirks, spread across different regions, and influenced vaccine strategies in subtle but important ways. Together, they defined the global COVID‑19 landscape in 2025.

Why 2025 Was Different?

By now, much of the world had entered what some called the “pandemic fatigue” era. Vaccines were widely available, booster campaigns had become routine, and governments were reluctant to re‑impose strict restrictions. Yet SARS‑CoV‑2 continued to adapt.

The six lineages that rose to prominence this year were not necessarily more severe, but they were smarter in evolutionary terms, finding ways to spread despite widespread immunity.

The World Health Organisation (WHO) noted that several of these lineages carried mutations associated with immune escape, meaning they could partially evade antibodies generated by vaccines or prior infection.

Others were products of recombination, where two different viral strains infected the same person and exchanged genetic material, creating hybrids with new properties.

The Six Significant Lineages

1. XEC: A recombinant variant first detected in Germany, later dominant across Europe and South Africa.
2. KP.3.1.1: Studied extensively for its spike protein changes, with structural insights making it adaptable
3. LP.8.1: A descendant of KP.1.1.3, notable for its rapid spread in the UK and Australia.
4. NB.1.8.1 (“Nimbus”): Emerging from recombinant lineage XDV.1.5.1, it drew attention for its immune escape potential.
5. XFG (“Stratus”): A recombinant of LF.7 and LP.8.1.2, carrying four spike mutations linked to immune evasion.
6. BA.3.2 (“Cicada” -Special mention): A heavily mutated lineage with diverse symptoms, first identified in South Africa.

Each lineage tells a story about how SARS‑CoV‑2 continues to adapt; sometimes through gradual mutation, sometimes through sudden recombination, and always under the pressure of human immunity.

Global Impact

The six lineages did not spread evenly. LP.8.1 dominated in the UK, NB.1.8.1 surged in Asia, XEC became widespread in Europe, BA.3.2 spread across Africa, KP.3.1.1 was studied in South Africa before going global, and XFG made headlines in India.

This patchwork of dominance reflected both the regional dynamics of transmission and the global interconnectedness of travel and trade.

Setting the Stage

As we move through each lineage in detail, the goal is not just to catalogue mutations but to understand what they meant for people’s lives. From hospital wards to airports, from vaccine clinics to family gatherings, these variants shaped the rhythm of 2025.

They remind us that the pandemic is not a single event but an ongoing negotiation between human society and viral evolution.

XEC and XEC* (* means descendants): The Mosaic Variant

When scientists first sequenced XEC, they noticed something unusual; its genome resembled a patchwork of varying genetic material. Rather than following the typical clean branching pattern of descent, XEC contained fragments from two distinct Omicron lineages. Offspring like this are known as recombinants.

In evolutionary virology, recombination is the equivalent of shuffling a deck of cards, where sometimes the outcome is trivial, but occasionally, it produces a hand that plays better than either parent.

The story of a hybrid

XEC was first detected in Germany in June 2024, derived from the JN.1 descendant lineages KS.1.1 and KP.3.3. By September 2024, the WHO had designated it a Variant Under Monitoring. By January 2025, XEC accounted for nearly 45% of global sequences submitted to GISAID, making it the most common Omicron subvariant at the time.

Why recombination matters

Mutation is the slow grind of evolution, but recombination is its shortcut. Instead of waiting for gradual changes, recombination can instantly assemble a new combination of traits.

For XEC, that meant inheriting immune‑escape mutations from one parent and transmission‑friendly tweaks from the other. It didn’t reinvent COVID‑19 symptoms, but it did make reinfections more likely, even in populations with high immunity.

Symptoms and lived experience

Clinically, XEC looked familiar due to common symptoms of sore throat, congestion, cough, fatigue, and fever. What stood out was not severity but persistence. People reported feeling “dragged out” longer than with other Omicron infections, with fatigue lingering for days.

Reinfections became a talking point in workplaces and schools, though not life‑threatening, but disruptive enough to keep absenteeism high.

Vaccines in the recombinant era

Vaccines continued to lessen the worst outcomes. Breakthrough infections were common, but hospitalisation rates stayed manageable. Immunologists pointed out that while recombination can shuffle antibody‑sensitive regions, T‑cell immunity recognises broader viral features, keeping protection against severe disease intact.

Global footprint

XEC spread widely across Europe and South Africa, becoming the dominant lineage in early 2025. In South Africa, prevalence rose from 0.5% in September 2024 to nearly 16% by December 2024 (See Figure 2). Wastewater surveillance picked it up early, and sequencing confirmed its steady climb.

Lessons from XEC

XEC’s story is less about clinical shock and more about evolutionary creativity. It reminded virologists that SARS‑CoV‑2 doesn’t just mutate, it recombines. That ability to reshuffle its genome means the virus can continue to surprise us, even when individual mutations appear familiar.

For public health the lesson was clear that surveillance must remain vigilant because recombinants can emerge quietly and spread steadily before anyone notices.

2. KP.3.1.1 and KP.3.1.1* : The Quiet Dominator

KP.3.1.1 didn’t arrive with fireworks. It wasn’t a sudden shock like Delta in 2021 or JN.1 in late 2023. Instead, it crept into the global landscape with quiet persistence, gradually outcompeting its cousins.

By late 2024, sequencing data from South Africa and Europe showed KP.3.1.1 steadily rising in prevalence. By early 2025, it had become one of the most studied lineages, with detailed structural insights through cryoelectron microscopy imaging (See Figure 3).

Mutational profile

KP.3.1.1 carried spike protein mutations that sharpened its binding affinity to ACE2 receptors. These included changes in the receptor‑binding domain (RBD) that subtly improved transmissibility and reduced neutralisation by antibodies.

Outside the spike, smaller tweaks in non‑structural proteins may have stabilised replication efficiency.

Unlike recombinant lineages such as XEC, KP.3.1.1 was a straight descendant, showing the power of incremental evolution. Its advantage wasn’t dramatic immune escape, but a steady accumulation of mutations that made it fitter in everyday transmission.

Symptoms and clinical impact

Clinically, KP.3.1.1 stayed within the Omicron pattern of sore throat, cough, congestion, fatigue, and fever. Loss of smell and taste remained rare compared to earlier pandemic waves.

Reinfections were common, often milder, but disruptive to workplaces and schools. Hospital clusters suggested higher reinfection rates among healthcare workers, though severity remained low.

Vaccines and immunity

Vaccines continued to protect against severe disease, even as KP.3.1.1 demonstrated immune escape at the antibody level. Breakthrough infections were expected, but boosters restored short‑term protection. T‑cell immunity remained robust, anchoring protection against hospitalisation.

Hybrid immunity, vaccination plus prior infection, provided the strongest shield.

Regions and timing

KP.3.1.1 was first noted in South Africa in late 2024, where it displaced other Omicron sublineages. By early 2025, it had spread globally, with peaks aligning to winter in the Northern Hemisphere. Wastewater surveillance in the US and Europe picked it up weeks before hospital admissions rose.

Public health lessons

KP.3.1.1 reinforced the importance of seemingly boring but effective interventions, including ventilation, boosters for vulnerable groups, and consistent surveillance. Governments avoided sweeping restrictions, instead focusing on targeted campaigns. Hospitals leaned on rapid triage and flexible staffing to manage rolling waves.

Looking ahead

KP.3.1.1’s descendants are likely to continue the “quiet advantage” strategy of incremental mutations, steady spread and familiar symptoms. The watch items are small but telling with subtle RBD changes, reinfection rates and any signs of recombination that could accelerate its trajectory.

3. LP.8.1 and LP.8.1.*: The Variant That Learned to Linger

LP.8.1 was not a headline‑grabbing recombinant like XEC, nor did it carry the dramatic symptom diversity of BA.3.2. Instead, it was a descendant lineage that demonstrated the quiet power of incremental evolution.

Emerging from KP.1.1.3, itself a branch of JN.1, LP.8.1 rose steadily through late 2024 and early 2025, eventually becoming dominant in regions such as the United Kingdom and Australia.

Its story is one of persistence rather than shock, showing how SARS‑CoV‑2 continues to adapt in subtle ways that nonetheless reshape the pandemic landscape.

A lineage built on persistence

The significance of LP.8.1 lay in its ability to outcompete other Omicron descendants without relying on recombination. Its spike protein carried mutations in the receptor‑binding domain that improved binding to ACE2 receptors, giving it a transmission edge.

These changes were not revolutionary, but they were enough to tilt the balance in crowded winter households in the UK and during summer travel in Australia.

Genomic surveillance confirmed its steady rise, with the UK Health Security Agency reporting that LP.8.1 accounted for more than half of sequenced cases by spring 2025.

Symptoms and reinfections

Clinically, LP.8.1 did not introduce a radically new symptom profile. It remained consistent with the Omicron pattern, which includes sore throat, cough, congestion, fatigue, and fever. Reinfections were common, often milder than first infections, but disruptive nonetheless.

Anecdotal reports suggested slightly higher reinfection rates compared to contemporaneous lineages, though severity remained low. Loss of smell and taste, once a hallmark of earlier variants, was rare, and most cases resolved within a week, although lingering fatigue was reported in some reinfections.

Vaccines and immunity

Vaccines continued to protect against severe disease, even as LP.8.1 demonstrated immune escape at the antibody level. Breakthrough infections were expected, but boosters restored a meaningful layer of protection.

Hybrid immunity, the combination of vaccination and prior infection, remained the strongest shield, providing broader and more durable protection than either alone.

Updated boosters offered short‑term protection against symptomatic disease, particularly valuable for older adults and those with underlying conditions. At the same time, T‑cell immunity continued to anchor protection against hospitalisation.

Regions and timing

LP.8.1’s dominance was most visible in the UK and Australia. In the UK, its rise coincided with the winter season, amplifying transmission in indoor settings.

In Australia, it surged during the summer travel season, contributing to rolling waves of reinfections that disrupted workplaces and schools.

Globally, LP.8.1 was detected across multiple continents, though its dominance was regional rather than universal. Wastewater surveillance and genomic sequencing caught its rise early, underscoring the importance of continuous monitoring.

Lessons from LP.8.1

The rise of LP.8.1 reinforced the importance of routine surveillance and booster campaigns. Its spread was not accompanied by increased severity, but reinfections strained workplaces and schools, highlighting the need for flexible sick‑leave policies and ventilation improvements in public spaces.

Public health messaging emphasised the basics, which included staying home when sick, protecting vulnerable populations and maintaining genomic sequencing to detect subtle shifts before they accumulate into dominance.

Looking ahead

LP.8.1’s descendants are likely to continue the incremental adaptation strategy, which includes small tweaks in the spike protein, steady transmission, and a familiar clinical presentation.

The main things to keep an eye on are reinfection rates, breakthrough cases, and any signs of recombination that could speed up its spread. Staying alert doesn’t mean panicking, but it does mean staying consistent, since LP.8.1 proved that even low‑profile variants can change the course of the pandemic.

NB.1.8.1 and NB.1.8.1* : The Variant Under Watch

NB.1.8.1, nicknamed ‘Nimbus’, entered the global conversation in early 2025 with a different kind of spotlight. Unlike LP.8.1, which rose through incremental mutation, NB.1.8.1 was quickly flagged by scientists and public health agencies as a Variant Under Monitoring.

Its roots traced back to a recombinant lineage known as XDV.1.5.1, and from the moment it appeared, the question was not whether it would dominate, but whether it would quietly reshape transmission patterns across continents.

A monitored lineage

The World Health Organisation added NB.1.8.1 to its monitoring list in May 2025, citing its potential for immune escape and its rapid spread across Asia and Europe.

The National Institute for Communicable Diseases (NICD) in South Africa also issued a technical brief, noting that NB.1.8.1 carried spike mutations associated with reduced neutralisation.

These early warnings framed NB.1.8.1 as a lineage worth watching, even before it reached high prevalence.

Genetic features

NB.1.8.1’s significance lay in its recombinant origin. By inheriting genetic material from XDV.1.5.1, it carried a combination of mutations that gave it a modest but noticeable growth advantage.

The spike protein showed changes in the receptor‑binding domain that reduced antibody recognition, while other tweaks suggested improved replication efficiency.

This was not a radical reinvention, but a careful reshuffling of traits that made reinfections more likely in populations with waning immunity.

Symptoms and clinical presentation

Clinically, NB.1.8.1 did not diverge dramatically from the Omicron symptom profile. Patients reported sore throat, cough, congestion, fatigue, and fever, with most cases resolving within a week.

Reinfections were common, often milder, but disruptive to daily life. Importantly, there was no evidence of increased severity compared to other Omicron lineages, a finding echoed in multiple surveillance reports.

Vaccines and immunity

Vaccines continued to protect against severe disease, even as NB.1.8.1 demonstrated reduced neutralisation in laboratory studies. Breakthrough infections were expected, but boosters restored short‑term protection.

Hybrid immunity remained the strongest defence, and T‑cell responses provided durable protection against hospitalisation. The WHO emphasised that while NB.1.8.1 showed immune escape, vaccines still did their most important job of preventing severe illness and death.

Regions and timing

NB.1.8.1 was first detected in Asia, with clusters reported in India and Southeast Asia. By mid‑2025, it had spread to Europe and the United States, appearing in GISAID submissions from multiple countries. Its prevalence rose steadily, though it never reached the dominance of XEC or LP.8.1.

Instead, it became part of the patchwork of lineages shaping the global COVID‑19 landscape. Surveillance data suggested its peak occurred in mid‑2025, coinciding with seasonal surges in the Northern Hemisphere.

Lessons from NB.1.8.1

NB.1.8.1’s story is less about clinical severity and more about the importance of vigilance. Its recombinant origin and immune escape potential made it a variant worth monitoring, even if its impact was moderate.

The rise of NB.1.8.1 underscored the value of wastewater monitoring, genomic sequencing, and timely booster campaigns.

5. XFG and descendants: Stratus Rising

By mid‑2025, one name began to dominate sequencing reports and public health briefings which was XFG, nicknamed ‘Stratus’. Unlike earlier lineages that rose quietly or regionally, XFG established itself as the current dominant Omicron variant, spreading across continents and reshaping the global COVID‑19 narrative.

Its story is not just about mutations, but about how recombination continues to drive the virus’s adaptability.

A recombinant with a clear edge

XFG emerged as a recombinant of LF.7 and LP.8.1.2, two lineages already circulating widely. Recombination gave it a mosaic genome, and within that mosaic were four spike mutations that mattered: His445Arg, Asn487Asp, Gln493Glu, and Thr572Ile.

These changes were linked to immune evasion, allowing XFG to slip past antibodies generated by prior infection or vaccination. The World Health Organisation assessed its risk as low in terms of severity, but acknowledged its potential to spread quickly due to immune escape.

Symptoms and clinical presentation

Clinically, XFG did not reinvent the disease. Most infections presented with mild to moderate symptoms of fatigue, cough, sore throat, and congestion. Fever was common, but loss of smell and taste remained rare.

What stood out was the sheer number of reinfections. People who had already experienced Omicron infections found themselves sick again, often with milder courses but still disruptive to daily life.

In India, where XFG was first reported in clusters across Maharashtra and Tamil Nadu, doctors noted that patients often described the illness as “short but draining”.

Vaccines and immunity

Vaccines continued to protect against severe disease, even as XFG demonstrated immune escape at the antibody level. Breakthrough infections were common, but boosters restored short‑term protection.

Immunologists emphasised that T‑cell immunity remained robust, anchoring protection against hospitalisation and death. WHO reiterated that vaccines were still effective where it mattered most, even if symptomatic infections were harder to prevent.

Global footprint

XFG’s rise was rapid. In India, it became the dominant lineage by mid‑2025, accounting for the majority of sequenced cases. From there, it spread across Asia, Europe, and North America, displacing other Omicron descendants.

Wastewater surveillance in multiple countries confirmed its prevalence, and sequencing data showed its steady climb. Unlike LP.8.1, which dominated regionally, XFG established a global presence, making it the defining variant of 2025.

Lessons from XFG

XFG’s dominance underscores the importance of recombination in SARS‑CoV‑2 evolution. It did not cause more severe disease, but its ability to evade immunity made it a disruptive force.

For public health, the lesson was clear that variants do not need to be more deadly to matter. Reinfections, absenteeism, and the strain on healthcare systems come not only from severity but from sheer numbers. Surveillance, booster campaigns, and ventilation improvements remain the pillars of response.

6. BA.3.2 and BA.3.2*: Cicada and the Long Tail

BA.3.2, nicknamed Cicada, first appeared in South Africa in late 2024 and quickly drew attention for its unusually heavy mutational load across the spike protein.

Unlike recombinants such as XEC or XFG, Cicada is a descendant lineage shaped by gradual mutation, but its density of changes in the receptor‑binding domain (RBD) and N‑terminal domain (NTD) of Spike protein suggested strong immune pressure, perhaps in a chronically infected host where the virus had time to accumulate alterations.

Clinically, BA.3.2 produced the similar Omicron symptoms of sore throat, cough, congestion, fatigue, and fever, but with a wider range of experiences. Patients reported night sweats, fainting spells, and prolonged fatigue, earning Cicada its nickname as a variant that produced a “chorus” of overlapping, sometimes unexpected symptoms.

In my own article titled, ‘Omicron BA.3: The Undead Variant Poised to Be the Chassis for Future Unstoppable Superspreaders’, I argued that BA.3’s resilience made it a kind of “undead chassis”, a framework upon which future variants could build. BA.3.2 exemplified that idea, showing how a lineage could persist and diversify even without global dominance.

By late 2025, Cicada’s descendants, collectively referred to as BA.3.2* and RE* began to rise in prevalence in several regions. Genomic surveys in Europe, Africa, and Asia reported clusters of sublineages carrying additional spike mutations that further reduced antibody neutralisation.

These descendants did not reinvent the clinical picture but fine‑tuned immune escape, making reinfections more common in populations with waning immunity. Wastewater monitoring in Germany and South Africa confirmed their growth before hospital admissions rose

Clinically, BA.3.2* infections resembled Omicron’s familiar profile, but anecdotal reports described more diverse symptom sets, echoing Cicada’s chorus. Vaccines continued to protect against severe disease, with boosters restoring short‑term protection, while hybrid immunity remained the strongest defence.

Together, BA.3.2 and its descendants illustrate the long tail of Omicron evolution. Even when a lineage does not dominate globally, its descendants can resurface with new mutations that give them regional advantages.

Cicada’s persistence and diversification show that SARS‑CoV‑2 evolution is not only about headline variants like XFG, but also about quieter families that continue to adapt, disrupt daily life through reinfections, and provide the genetic scaffolding for future superspreaders.

Conclusion: The Shape of Viral Evolution in 2025

Looking back across the six lineages we have explored, KP.3.1.1, XEC, LP.8.1, NB.1.8.1, BA.3.2 (Cicada), and XFG (Stratus), what emerges is not a single narrative of dominance, but a tapestry of evolutionary strategies.

Each lineage tells a different story about how SARS‑CoV‑2 adapts, persists, and reshapes the pandemic’s rhythm.

KP.3.1.1 illustrated the power of incremental mutation, a “quiet dominator” that spread steadily without dramatic recombination. XEC reminded us of the mosaic potential of recombination, showing how genomes can be reshuffled to produce hybrids with subtle but effective advantages.

LP.8.1 demonstrated the regional dominance of descendants, rising to prominence in the UK and Australia through persistence rather than novelty. NB.1.8.1, flagged as a Variant Under Monitoring, underscored the importance of vigilance even when clinical severity remains low.

BA.3.2, nicknamed Cicada, revealed the long tail of Omicron evolution, producing diverse symptoms and spawning descendants (BA.3.2* and RE*) that are now resurging in parts of Europe, Africa, and Asia.

Finally, XFG, the recombinant Stratus, showed how a lineage can achieve global dominance by combining immune escape mutations with efficient transmission, spawning its own family of descendants that continue to fine‑tune its success.

Together, these lineages highlight three recurring themes. First, evolution is modular, i.e., SARS‑CoV‑2 adapts through both gradual mutation and recombination, with each strategy producing viable outcomes.

Second, dominance is contextual, some lineages rise globally, others regionally, and still others persist as reservoirs of genetic diversity.

Third, impact is not only about severity but reinfections, absenteeism, and disruption to daily life matter just as much as hospitalisation rates.

For public health, the lesson of 2025 is continuity. Surveillance through genomic sequencing and wastewater monitoring remains the cornerstone of preparedness.

Vaccines continue to do their most important job by protecting against severe disease, even as immune escape produces breakthrough infections.

Simple interventions such as ventilation, sick‑leave flexibility and targeted booster campaigns are what turn waves into ripples.

In my own reflections, including the article Omicron BA.3: The Undead Variant Poised to Be the Chassis for Future Unstoppable Superspreaders, I argued that BA.3 represented a kind of evolutionary chassis.

The events of 2025 confirm the idea that, whether through recombination or mutation, SARS‑CoV‑2 continues to build upon existing frameworks, producing lineages that are not necessarily more deadly but more adaptable.

As we move into 2026, the challenge is not to predict a single “next big variant,” but to understand the ecosystem of lineages that together shape the pandemic. The virus has shown us that it thrives on diversity.

Our response must do the same of layered, flexible, and prepared for both the quiet dominators and the sudden mosaics that will inevitably emerge.

Stay Safe!

From Cicada to Stratus: 6 COViD Variants That Shaped How We Lived in 2025 was originally published in Microbial Instincts on Medium, where people are continuing the conversation by highlighting and responding to this story.

From viral mutations to human resilience, a scientist’s reflection on order, chaos, and meaning.

I still remember the first time I watched ice melt in a glass of water. It was a simple moment, but it carried a profound lesson. The solid structure dissolved into liquid, order gave way to disorder, and yet nothing was lost… only transformed. That was entropy at work.

At the time, I thought of it as a scientific curiosity. Today, after years of studying viruses, mutations, and the relentless adaptability of life, I see entropy as something far greater. It is not just a law of physics. It is a philosophy of existence, a reminder of impermanence, and a silent teacher urging us to act with urgency and resilience.

Entropy in Equations

Entropy is often defined as a measure of disorder, but that definition is misleading. It is not chaos in the everyday sense. It is the dispersal of energy, the tendency of systems to move towards equilibrium. In thermodynamics, we write it as:

ΔS = Q / T

where Delta S is the change in entropy, Q is the heat transferred, and T is the absolute temperature.

In the Gibbs free energy equation,

ΔG = ΔH — TΔS

where ΔG is change in Gibbs free energy, ΔH is change in enthalpy (heat exchange at constant pressure), T is absolute temperature, ΔS is change in entropy

Entropy appears alongside enthalpy, determining whether a process is spontaneous.

These equations are precise, but they also convey a deeper truth: entropy serves as the universe’s reminder that change is inevitable. Nothing remains perfectly ordered forever; systems evolve, structures dissolve, and equilibrium eventually prevails.

Viruses as Masters of Entropy

In virology, entropy is a tangible concept not just an abstraction. It is visible in every replication cycle of HIV, in every recombination event of SARS-CoV-2, in every mutation that reshapes the evolutionary landscape of viruses.

HIV is notorious for its high mutation rate because each replication introduces small changes, creating a swarm of variants known as a quasispecies.

Most of these mutations are harmful, leading to defective proteins or non-functional viruses. However, some mutations can provide advantages, such as drug resistance, immune evasion, or enhanced replication. Entropy ensures that the viral population is constantly changing, spreading various possibilities across the sequence space and allowing for survival even under intense selective pressure.

SARS-CoV-2, the virus behind COVID-19, offers another vivid example. Its evolution has been marked by mutations in the spike protein, particularly in the receptor-binding domain, and by recombination events that reshuffle its genome.

The emergence of variants such as Delta and Omicron was not random chaos but entropy-driven exploration filtered by natural selection. Entropy provided the raw material that included mutations and recombinations. Selection favoured those that improved transmissibility or immune escape.

The result was a dynamic evolutionary landscape, a reminder that entropy and order are not enemies but partners in shaping life.

Philosophy — Fragility of Order

Philosophically, entropy challenges our desire for permanence. We build structures, institutions, and even identities as if they could resist the tide of time. Yet entropy whispers that all things dissolve. Ice melts, stars burn out, civilisations rise and fall.

In science, this inevitability is expressed through Gibbs free energy: progress happens when energy and openness to change align.

Philosophically, it tells us that transformation requires both the fuel of enthalpy and the willingness to embrace entropy. It is not enough to have energy; there must also be acceptance of disorder. Only then does change become possible.

Entropy has become a metaphor for resilience in my understanding. Studying viruses has shown me that entropy is not the enemy of order; rather, it is a necessary condition for adaptation. Without entropy, there would be no mutations, no recombination, and no evolution. Life itself would become stagnant.

Entropy ensures diversity, and diversity is the raw material of survival. HIV survives because its entropy-driven mutations allow it to escape drugs and immune responses. SARS-CoV-2 continues to circulate because entropy provides new variants that challenge our defences. Entropy is the silent force that keeps life dynamic, unpredictable, and creative.

But entropy is not confined to biology. It is present in human existence. In learning, entropy manifests as forgetting. Knowledge resists entropy only if we continually invest energy in study and practice.

In relationships, entropy manifests as distance. Bonds dissolve unless we nurture them with attention and care. In creativity, entropy manifests as silence. Ideas fade unless we express them, share them, and give them form.

Entropy as a Teacher

Entropy teaches us that maintenance is as important as creation. To keep order, we must continually invest energy, just as living systems must continually fight against equilibrium.

Entropy also teaches humility. The universe does not care about human values. Stars will burn out, galaxies will disperse, and eventually, if the heat death scenario holds true, all structures will dissolve into maximum entropy. For some, this is terrifying nihilism. For others, it is liberating.

If the universe is indifferent, meaning is not given; it must be created. Entropy reminds us that permanence is an illusion, but purpose is a conscious choice. We cannot stop the tide of disorder, but we can shape our response to it.

We can choose to build, to love, to create, knowing that entropy will one day reclaim it all. That knowledge does not diminish our efforts; it makes them more urgent, more precious.

In virology, entropy is both a challenge and an opportunity. Drug resistance in HIV is a direct consequence of entropy. Mutations accumulate, and some confer resistance to antiretroviral therapy. This forces us to innovate, to develop new drugs, to design combination therapies that make it harder for entropy to win.

In COVID-19, entropy-driven mutations challenge vaccines, requiring updates and boosters. Yet entropy also provides opportunities for science. By studying the patterns of mutation and recombination, we learn how viruses evolve, how they adapt, and how we can anticipate their moves. Entropy becomes a teacher, guiding us towards deeper understanding and more effective interventions.

Reflecting on entropy has also shaped how I see my own life. As a scientist, I know that entropy will eventually dissolve my work, my words, my existence. But that does not make the effort meaningless. On the contrary, it makes it more meaningful. Every paper written, every visualisation created, every student taught is an act of defiance against entropy. It is a way of investing energy to create order, however temporary.

Entropy reminds me that time is precious, that procrastination is costly, that the opportunity to act is fleeting. It pushes me to write now, to teach now, to create now, because tomorrow entropy will have advanced, and the chance may be gone.

Entropy connects science and philosophy in a way few concepts do. It is a physical law, a measurable property, a mathematical quantity. But it is also a metaphor for impermanence, a symbol of mortality, a reminder of humility. It bridges the laboratory and the human heart.

It tells us that the same law that governs the melting of ice also governs the fading of memories, the dissolution of relationships, and the evolution of viruses. Entropy is universal, impartial, and inevitable. Yet within its inevitability lies the possibility of meaning. We cannot stop entropy, but we can choose how to live within its flow.

In conclusion..

Entropy is not just a scientific concept. It is a philosophy of life. It tells us that change is inevitable, that permanence is an illusion, and that resilience is essential. In virology, entropy drives mutation, recombination, and evolution, shaping the survival of viruses like HIV and SARS-CoV-2. In human life, entropy challenges us to maintain knowledge, nurture relationships, and express creativity. In philosophy, entropy reminds us that meaning must be created, not found. The universe may not care, but we can. Entropy is the silent teacher urging us to grow, to adapt, and to live fully before equilibrium reclaims us.

Entropy — The Silent Teacher of Life was originally published in Predict on Medium, where people are continuing the conversation by highlighting and responding to this story.

What if nature, or the universe as a whole, is not concerned about us as individuals but has established several laws, or algorithms, to govern matter?

I mean to think that the universe would care to bring me back down to the ground when I jump would be an overinflated ego. Perhaps, just like in coding, it has created an algorithm that ensures whatever goes up comes back down at a certain speed based on its size, shape, or location.

It would be similar to the following Python function:

def gravity(object):
    if object.elevated:
        if object.mass >= 10:
            if object.shape == "round":
                if object.location == "Earth":
                    object.gravity_acceleration = "10 m/s^2"

Of course, the Universe’s algorithm would be orders of magnitude more complex than this, but perhaps in a structure as illustrated above.

The laws would be so irrefutable that they must be followed; there won't be an object or creature, whether good or bad, that is too special not to comply.

What inspired me to write this piece is that over the past few days, I have noticed and decided to study the energy that surrounds people and other objects. I chose the word "notice" to emphasise that I used my human senses to explore the energies present in and around matter.

Not scientific because it's not reproducible, I know, but…

In my life and perhaps in the lives of most people, there are individuals with whom I connect, and others with whom I struggle to find that connection. Similarly, some places feel serene, while others feel hostile, often for no apparent reason.

It could be something as simple as an ice cream flavour that feels right or hypnotises you to choose it.

Why is that?

Reading an object’s energies is like interpreting people’s facial expressions. The key difference is that energy cannot be faked like facial expressions; it just can’t lie.

Often, we as humans are aware of what is truly happening in a situation, but we choose to ignore it because we want to believe something different. We may desire to alter reality, perhaps to escape from pain or to seek pleasure.

It's understandable that in our advanced world today, with all this technology, we often forget that our bodies have not changed anatomically or behaviorally since the Stone Age, when we were hunting and gathering.

We have created advanced machines that can even outsmart us. This gives us a sense of distinct superiority over other animals and even inanimate objects, which leads to a cognitive gap. If we consider ourselves at the top of the natural hierarchy, we must ask, is there something else above us?

To address this gap, many cultures have conceptualised a higher power or deity, an entity superior to humans. The rest of the story is for another day.

My point is that the energy we feel, often referred to as intuition, instinct, insight, or gut feeling, actually comes from our connection with our surroundings. Many people have neglected this sense because life today is much safer physically than it was in the jungles and forests.

Ralpho Waldo Emmeson in his Essay called Compensation said, ‘The law of nature is: Do the thing, and you shall have the power,’ he then concludes, ‘but they who do not the thing have not the power.’

If you feel the life you are living is not what you expect of yourself, or that it could be better than this, study your life with scrutiny. ‘Know Thyself’(reminds me of Neo and the Oracle), and see which ‘thing’ you are not doing that keeps you from having the power. What is your yoke?

Eliminate every ‘thing’ that works against your true power. Dare to explore how much power you can wield, how you can change the world into a better place. There is no price greater, no sacrifice more tragic, than prostituting your ultimate power.

Many people are afraid of their true selves — afraid it might be too dark to be socially acceptable. But guess what? It’s your energy. Human nature tends to label energies as good or bad, but like I said in my opening paragraphs, Nature doesn’t care. All it wants is authenticity. And when you honour that, you’ve inputted the correct variables and values to run the fulfilment algorithm. EXPLORE!

Of course, there are numerous algorithms in Nature. A life well-explored is a life well lived.

When people pressure you to choose something that isn't you, it's usually for their own benefit; they are being selfish. So, why not be selfish too? Choose who matters to you.

My point is: follow the energy. At first, it might feel heavy, but it will grow lighter to bear. Because in those moments, you’re swimming in your compatible energy. You’ll feel awake and alive, not just merely existing.

If you should choose anything in life, choose your purpose — the one that feels like a lighter burden to carry in the long run. Don’t look to the side for ideas from others pursuing their own paths. It’s not a race.

I know this piece is composed of many seemingly unrelated fields of study, but let me weave it together conclusively:

The energy exuded by objects is an algorithm written by Nature (or, if you prefer, the Universe) to prompt movement toward or away from them. No matter how prayerful, hopeful, or motivated you are, no matter what word you use to force it to be yours or not yours, the energy speaks for itself. Follow the Energy!

What if Nature works with algorithms too? was originally published in Predict on Medium, where people are continuing the conversation by highlighting and responding to this story.

When you first meet lenacapavir (LEN) in the literature, it feels different. Most antiretrovirals block enzymes; lenacapavir attacks the virus’s architectural core. That choice alters the logic of inhibition. Instead of blocking a single catalytic step, a capsid inhibitor disrupts a structural assembly that the virus utilises at multiple points in its life cycle.

The conceptual shift is exciting and important for clinicians, program designers, and communities. Lenacapavir is not only a novel mechanism but is also delivered as a long-acting subcutaneous depot. This form of treatment alters our perspective on adherence, healthcare delivery, and the ethical trade-offs involved.

I studied papers on structural biology, pharmacokinetics, and clinical summaries to translate the findings into language that is easy to understand.

This article explains the fundamental science, connects it to clinical practices and programmatic realities, and discusses the implications of the evidence for promoting equitable implementation. It is also an in-depth scientific version to my last article on the geopolitics of HIV drug administration in African countries.

Additionally, though there are 2 HIV strains, this article will focus on HIV-1, which is the globally dominant strain responsible for the majority of infections and is generally more transmissible.

HIV-2 is largely concentrated in West Africa and transmits less efficiently and leading to slower disease progression. Both viruses share the core biology of a lentiviral life cycle, use of CD4 and co-receptors for entry and enzymatic genes.

However, they differ in sequence, antigenicity and clinical behaviour enough that some diagnostics, drug susceptibilities and epidemiological patterns diverge.

Clinically, HIV-2 is less responsive to certain antiretrovirals developed against HIV-1 and requires careful laboratory confirmation and tailored treatment decisions; hence, understanding which virus a person carries matters for prognosis, monitoring and therapy.

1. Why gag → CA matters?

HIV-1 packages a positive-sense RNA genome of approximately 9.8 kilobases inside each virion. Whats does this mean?

Ribonucleic acid (RNA), particularly messenger RNA (mRNA), serves as an intermediary in the process of translating genetic information from deoxyribonucleic acid (DNA) into a polypeptide, which may or may not be a protein. This is the usual way of things in higher organisms like plants and animals.

When an RNA virus enters a host cell, it must integrate or recombine its genome with that of the host cell to produce its own proteins. Since DNA and RNA are different coding languages, the virus must reverse-transcribe its RNA into DNA to match the language used by the host cell. This is called a negative-sense RNA.

A positive-sense RNA (like HIV) functions like mRNA, as it does not need to convert to DNA; instead, it directly uses its RNA genome to produce proteins.

The Group-specific antigen (gag) open reading frame produces a polyprotein that, when cleaved by viral protease during maturation, yields matrix (MA), capsid (CA), nucleocapsid (NC) and p6 (C-terminal fragment).

CA is the structural unit that builds the capsid shell and is the direct molecular target of lenacapavir. CA is synthesised inside budding virions and self‑assembles after protease cleavage. Each CA monomer contains two folded domains, the N‑terminal domain (NTD) and the C‑terminal domain (CTD), connected by a flexible linker (see image below).

The N-terminal domain refers to the end of a polypeptide that has an amino group (containing an N atom), while the C-terminal domain denotes the end of a polypeptide chain that includes a carboxylic acid group (containing a C atom).

The NTD forms part of the hexamer (6 CA monomers) central pore; the CTD mediates dimerisation and links hexamers into the lattice. Because CA forms the mechanical scaffold that both protects and times genome release, many of its interface residues are functionally constrained.

In other words, pockets used to hold the lattice together tend to be less tolerant to mutation, a practical advantage for small‑molecule targeting.
For readers who prefer a clinical perspective, this is why CA is a viable drug target and less susceptible to viral escape than some enzyme targets.

For implementers and communities, this means the drug's effects and risks will differ from those of traditional enzyme inhibitors.

2. Think of the capsid as engineered geometry

The capsid is a geodesic shell, a fullerene cone made from repeating CA units (shown above). The mature HIV capsid consists of approximately 250 hexamers and exactly 12 pentamers. Hexamers form the bulk of the lattice and create planar regions; pentamers (5 CA monomers) are positioned where high curvature is required to close the cone.

The precise placement of pentamers versus hexamers determines curvature, mechanical rigidity and the internal space the capsid protects.
CA–CA interactions are mediated across distinct interfaces, NTD–NTD contacts form hexamer cores, NTD–CTD interactions stabilise geometry, and CTD–CTD dimers stitch hexamers together.

The capsid is dynamic; it must be stable enough to protect reverse transcription yet flexible enough to uncoat at the right time, so the viral genome can enter the nucleus and integrate.

Host proteins, such as nucleoporins and transport adaptors like CPSF6, recognise specific surfaces on the capsid and facilitate its trafficking through the nuclear pore. Altering the geometry of the capsid affects the timing and quality of these molecular interactions.

3. What lenacapavir was designed to do, and why that matters?

Medicinal chemists developed lenacapavir to target a conserved hydrophobic pocket at CA intersubunit interfaces (see image below). The goals were straightforward. They aimed to ensure high affinity and prolonged residence time in the pocket.

Additionally, they wanted to maintain effectiveness across various HIV‑1 sequences. Lastly, achieving physicochemical properties suitable for a long-acting depot was also a key objective.

In vitro, lenacapavir exhibits exceptional potency, with inhibitory concentrations ranging from sub‑nanomolar to low‑nanomolar across various assay formats.

The drug exhibits a bimodal biological signature by inducing both early-stage defects, such as reverse transcription and nuclear import, along with late-stage defects that involve the assembly and formation of infectious virions.

Clinically, developers used an oral lead‑in to confirm tolerability and serum exposure, followed by a subcutaneous depot for maintenance. The depot releases the drug slowly, smoothing peaks and extending the apparent half‑life so that dosing intervals can be measured in months rather than days.

That is the transformative promise of the product, reducing daily decisions about pills and replacing them with periodic clinic encounters.

During a WHO webinar I attended in late August this year, concerns were raised about the programmatic burdens associated with depot introduction. These include scheduling injections, the need for cold chain management, and supply chain considerations.

Additionally, it is necessary to train staff on administration procedures and establish clear protocols for handling missed appointments.

These are not minor operational details; they significantly influence how populations experience the technology.

4. The core mechanism: why lenacapavir creates “structural chaos”?

This was the most surprising part when I read the structural and biochemical literature. The hydrophobic pocket that lenacapavir occupies exists in both hexamer and pentamer forms (see image below), but the local geometry and energetics vary.

The drug, therefore, has different effects depending on the oligomeric state where it binds:

i. In hexameric regions, lenacapavir strengthens intersubunit contacts and reduces lattice dynamics, resulting in hyperstable hexamers.

ii. In regions that must form pentamers to close the cone, the drug interferes with pentamer nucleation or incorporation; pentamer formation is impaired.

The net effect is an imbalance where stabilised hexamer patches are adjacent to areas that cannot properly curve or close. That juxtaposition of hyperstable and defective regions is what structural biologists and biochemical assays describe as “structural chaos.”

Why does that matter?

Since the capsid’s function relies on coordinated global behaviour, it involves timely uncoating, the exposure of host-factor binding epitopes, and proper encapsidation during virion assembly.

If some regions refuse to uncoat while others are incomplete, reverse transcription stalls, nuclear import is derailed, and newly formed virions are malformed and non-infectious.

In short, by shifting the hexamer: pentamer equilibrium, lenacapavir hits the virus at several life‑cycle points simultaneously.

5. The experimental evidence that links structure to function

Several orthogonal lines of evidence make the lenacapavir story convincing:

i. Structural snapshots

Cryo‑EM and crystallographic studies show lenacapavir nestled in the intersubunit pocket; pocket interactions differ subtly in hexameric versus pentameric contexts.

ii. Biophysical assays

Thermal‑shift and hydrogen–deuterium exchange experiments report increased stability for hexamer interfaces and altered flexibility in regions associated with pentamer nucleation.

iii. In vitro assembly studies

Experiments that reconstitute CA assembly show altered hexamer pentamer ratios and the formation of non-canonical oligomers when the drug is present.

iv. Cell biology and virology

Treated infected cells show early defects consistent with stalled reverse transcription and impaired nuclear import. When virions are produced in drug‑exposed cells, they often contain defective cores and lose infectivity, demonstrating late‑stage effects (See image below).

v. Resistance genetics

In vitro selection and early clinical monitoring have identified substitutions in or around the pocket and adjacent interfaces that reduce susceptibility, and many of these changes incur fitness costs consistent with the pocket sitting in a constrained region.

Taken together, these datasets form a chain of evidence from the atomic level to the organismal level: binding → altered oligomerisation → altered capsid mechanics → functional failure.

6. Resistance: pathways, trade‑offs and the danger of the tail

Resistance pathways fall into categories:

a) direct pocket substitutions that reduce affinity

b) allosteric/interface mutations that indirectly reshape the pocket

c) compensatory changes that restore lattice properties when initial resistance hits impose fitness costs.

Often, resistance is not a single leap but a stepwise trajectory. There is both good news and concerning news.

The good news is that the interfaces are functionally constrained; many single substitutions result in large fitness costs, making escape more difficult than for some enzyme inhibitors.

The concerning news is that some substitutions confer modest resistance while preserving near-wild-type replication, which are realistic threats in the real world. Compensatory mutations can then restore fitness while preserving resistance.

The pharmacokinetic (PK) context of the depot intensifies this problem. After the final injection, plasma concentrations decline slowly, a prolonged PK tail.

If that tail includes concentrations that are not fully suppressive, selection for modestly resistant variants is encouraged. Missed injections, supply interruptions or delayed re‑loads create practical windows where resistance selection can occur.

That is why program design must include bridging regimens, strong recall systems and molecular surveillance focused on CA.

7. Pharmacokinetics - the power and the Achilles’ heel of long‑acting depot dosing

The depot’s gradual absorption and tissue binding result in an extended apparent half-life, which allows for six-monthly dosing. Clinically, this is a significant adherence innovation that liberates individuals from daily pills and diminishes the daily visibility of treatment.

But the same pharmacology produces the PK tail, which is a period of waning and sometimes sub-therapeutic concentrations. This tail is the operational Achilles’ heel.

In the WHO webinar, it was highlighted that the duration of the tail varies across individuals and populations, influenced by factors like body composition, age, pregnancy status, and liver function.

In practice, clinics must plan for missed injections, which requires bridging oral therapy, implementing rapid re-loading protocols, or taking other mitigation steps to avoid long windows of partial suppression.

Providers should counsel patients about the importance of attending injection appointments and make contingency plans for unavoidable delays.

Programs should keep track of interactions with other medications that impact hepatic transporters or biliary excretion, as these pathways are important for the elimination of lenacapavir.

8. Clinical evidence: where the drug has shown value so far?

In studies involving participants with extensive prior antiretroviral exposure, particularly those with multidrug-resistant HIV and few remaining treatment options, lenacapavir, when added to an optimized background regimen, demonstrated significant efficacy.

For example, in the Phase 2/3 CAPELLA trial (NCT04150068), 72 participants with heavily treatment-experienced HIV were enrolled. After 14 days of functional monotherapy with lenacapavir, 88% of participants achieved ≥0.5 log₁₀ reduction in viral load compared to 17% in the placebo group. By Week 26, 81% of participants achieved viral suppression (<50 copies/mL), underscoring lenacapavir’s potency even in salvage therapy settings

While much of the early clinical data on lenacapavir focused on treatment in heavily treatment-experienced individuals, recent trials have expanded its role into HIV prevention.

In particular, the PURPOSE 1 and PURPOSE 2 Phase 3 trials evaluated lenacapavir as long-acting pre-exposure prophylaxis (PrEP). In PURPOSE 1, involving nearly 5,000 cisgender women in sub-Saharan Africa, zero HIV infections occurred in the lenacapavir arm, demonstrating 100% efficacy.

PURPOSE 2 extended this to cisgender men, transgender women, and nonbinary individuals across multiple regions, showing 99.9% efficacy. These results underscore lenacapavir’s potential to transform HIV prevention by offering biannual dosing, which addresses adherence barriers and reduces stigma associated with daily oral PrEP.

Safety profiles from trials are encouraging; systemic adverse events are uncommon, and injection‑site reactions are the most frequent local issue, usually mild to moderate.

Long‑term post‑marketing data, including in pregnancy and pediatric populations, are still needed.

A practical approach is to treat lenacapavir as a powerful tool, ideally used in combination strategies for treatment. For PrEP users, careful testing and robust linkage systems are crucial to minimise breakthrough infections and ensure they are quickly managed with combination therapy when they arise.

9. Implementation and equity: how communities will actually experience this drug

The promise of a six‑monthly injection is emotional and practical. For people who struggle with daily adherence, a twice‑yearly depot can reduce stigma, free mental energy and increase retention.

For programs, depot products can lower the logistical burden of monthly pill distribution, but equitable access is not automatic. Long‑acting products can create new dependencies on supply chains, clinic infrastructure, and trained personnel.

Without negotiated pricing, technology transfer or local manufacturing, many high‑burden settings could face limited supply or unaffordable costs. Communities must be engaged in decisions about procurement, delivery models and consent processes.

Models that integrate depot delivery into maternal health clinics, community outreach services or decentralised platforms may work better than centralised clinic models in many contexts.

Operationally, programs should plan for:
a) appointment recall systems

b) protocols for delayed injections and bridging oral therapy

c) training for staff on injection technique and counselling

d) Sentinel surveillance for CA resistance linked to clinical outcomes.

10. Where should research focus next?

From reading the literature and thinking about implementation, several priorities stand out:

a) Resolve high‑resolution structures of clinically observed resistant variants to inform next‑generation inhibitor design.

b) Quantify PK tails across pregnancy, BMI ranges, pediatric populations and other groups to define bridging strategies.

c) Test combination regimens that minimise resistance risk while preserving tolerability in both treatment and PrEP contexts.

d) Conduct implementation research on decentralised delivery, missed‑dose mitigation and community preferences in high‑burden settings.

e) Build molecular surveillance that integrates CA sequencing with phenotypic follow‑up and program metrics.

f) Study procurement, pricing and technology transfer models that expand manufacturing and reduce inequitable dependence.

These are scientific and social problems that require interdisciplinary teams including structural biologists, clinicians, pharmacologists, program designers, ethicists, and community leaders working together.

11. A human reflection

This story sits at the junction of molecular elegance and social reality. The molecule is clever as it turns the virus’ structural dependence against itself, creating multi‑stage inhibition that is both scientifically interesting and clinically useful.

The vision of fewer pills and less stigma is compelling. Yet the science also insists on humility: PK tails, resistance trade‑offs and operational requirements mean that success depends as much on systems and equity as on chemistry.

If you are a clinician, researcher, policy-maker, or advocate, your role is to combine this innovation with effective delivery systems and strong surveillance. If you are part of a community affected by HIV, it is essential that your voice helps shape the use of this tool, including when it is introduced, who will have priority access, and how consent and choice are guaranteed.

I researched, read and synthesised the literature so we could have a humane, evidence‑based conversation about how best to adopt lenacapavir.

If you’ve read this far, please support me by following and subscribing to my email to ensure you don’t miss these valuable insights.

Lenacapavir: A New Class of Drug Transforming the Future of HIV Prevention was originally published in Microbial Instincts on Medium, where people are continuing the conversation by highlighting and responding to this story.

“This is more than medicine, it’s a pathway to a safer, stronger, and healthier future.” — (U.S. Embassy, Zimbabwe), but it’s not just about the science; it’s about the approach and the intent.

Lenacapavir, marketed as Sunlenca, is an antiretroviral drug for HIV recommended for individuals with multidrug resistance who have limited treatment options.

At the end of September 2025, the US Embassy in Zimbabwe posted on X, congratulating Zimbabwe for being selected as one of the African countries to first receive the drug. While reading the comments, I noticed that many people were skeptical, questioning the timing and potential hidden agendas.

This inspired me to write a piece encouraging everyone to consider both perspectives. I included the scientific aspects of the drug and emphasized that, while it appears revolutionary on paper, science alone is not enough, there are political issues that still need to be addressed.

The Scientific Promise: What Makes Lenacapavir So Remarkable

As a science educator and bioinformatics researcher, I have spent years studying how viruses evolve, how drugs interact with biological systems, and how public health decisions ripple through communities.

When I read the announcement that Zimbabwe and nine other sub-Saharan African countries had been selected to pilot lenacapavir, I paused.

Lenacapavir is a unique antiretroviral treatment, as it is the first of its kind to function as an HIV capsid inhibitor. This means it specifically targets the HIV protein shell, known as the capsid, which is essential for the virus’s replication, assembly, and ability to infect new cells.

Most HIV drugs act after the virus has already entered host cells and begun replication. Lenacapavir intervenes earlier, disrupting multiple stages of the viral lifecycle as shown in the diagram below:

Understanding the HIV Capsid: The Viral Shell That Holds the Key

The HIV capsid is a cone-shaped structure made up of repeating units of the capsid protein (CA). It serves to protect the viral RNA and enzymes as they travel into the host cell and plays a crucial role in the uncoating process once the virus is inside. This shell is not only structural but also functional, dynamic, and essential for the survival of HIV.

Lenacapavir binds to specific sites on the capsid and interferes with its ability to disassemble properly. It also disrupts the formation of new capsids during viral assembly, leading to malformed, hyper-stable virus particles that cannot infect target cells.

A recent study shows that lenacapavir exerts opposing effects on the formation of CA pentamers and hexamers, key intermediates in capsid assembly. By stabilizing hexameric lattices and impairing pentamer formation, it causes structural chaos within the virus.

Lenacapavir operates through a dual-action mechanism that blocks both the entry and exit of viruses, giving it a unique potency. It has shown picomolar-level efficacy in laboratory tests and has produced impressive results in clinical trials.

In Phase 3 studies (see the table below), lenacapavir achieved a prevention rate of 100% in cisgender women and a 96% prevention rate in cisgender men who have sex with men, as well as in transgender and gender non-binary participants.

The Long-Acting Advantage — One Injection, Six Months of Protection

What truly distinguishes lenacapavir is its long-acting formulation. Given as a subcutaneous injection just twice a year, it provides sustained protection without the daily hassle of taking pills. This is revolutionary for individuals who struggle with adherence due to factors such as stigma, mobility issues, or simply forgetting.

Think about the possibilities:

1. Discreet protection for young women, sex workers, and LGBTQI individuals who face judgment or violence.
2. Reduced healthcare visits, easing pressure on overstretched systems and freeing up resources.
3. Improved adherence, especially in rural and underserved communities where access to clinics is limited.

This represents science at its finest — a single injection providing six months of reassurance.

Why Sub-Saharan Africa? The Rationale Behind the Rollout

To be fair, there are clear scientific and public health reasons why sub-Saharan African countries are being prioritised for lenacapavir’s rollout.

1. High HIV Burden

Sub-Saharan Africa accounts for nearly two-thirds of global HIV infections. Countries like South Africa, Nigeria, Kenya, and Zimbabwe have some of the highest prevalence rates in the world. Targeting prevention here could have the greatest impact on global incidence.

2. Established Infrastructure

Thanks to decades of work by PEPFAR, the Global Fund, and local ministries of health, many African countries have robust HIV testing, monitoring, and treatment systems in place. These systems are capable of supporting the logistics of a long-acting injectable.

3. Proven Community Engagement

African health workers, educators, and civil society groups have built trust and resilience in the face of stigma and misinformation. Community-led initiatives have shown that when people are empowered, they respond with commitment and care.

4. Diverse Populations for Research

From urban centers to rural clinics, the region offers varied demographics for studying real-world effectiveness. This diversity is scientifically valuable and ethically necessary.

In summary, Africa is not just vulnerable but also prepared, which is why it has been chosen. Studies show that the drug has been tested on various global ethnicities. However, being prepared doesn’t eliminate risks, and gratitude shouldn’t suppress essential questions.

The Politics: Why Sympathy Now, When So Much Else Is Being Denied?

I cannot ignore the contradictions.

Why should the United States extend its medical generosity to sub-Saharan Africa now, especially when African scholars are being denied visas, funding for African-led research is being cut, NGOs are shrinking, and the global narrative continues to depict us as hypersexual, intellectually inferior, and perpetually dependent?

I’ve witnessed these stereotypes myself…the jokes, the dismissals, the assumptions. That Africans are overly sexual, that we can’t follow instructions, that we’re in constant need of saving. And now, out of nowhere, a miracle drug is being offered to us?

Is this compassion or containment?

Is lenacapavir a gift — or a geopolitical gesture?

I want to believe in the science. I do. But I also know that science does not arrive in a vacuum. It is shaped by politics, power, and history.

The Sub-Saharan Dilemma: Between Gratitude and Grievance

Sub-Saharan Africa is home to brilliant scientists, resilient health systems, and youth-led movements that challenge the status quo. Yet we are rarely treated as equals.

If lenacapavir is to succeed, it must be more than a pilot project. It must be rooted in:

1. Local Leadership

African researchers, clinicians, and educators must be at the forefront of implementation, monitoring, and education. We must not be passive recipients, we must be active architects.

2. Transparent Funding Models

We need clarity on who pays, for how long, and under what conditions. Will lenacapavir remain free?

Will governments be expected to shoulder costs once donor enthusiasm fades?

3. Respectful Engagement

The rollout must break down stereotypes and prioritize African dignity. It must acknowledge that our communities are not merely places of disease — they are also places of resilience, innovation, and hope.

“We must not confuse access with autonomy. A drug is not empowerment unless it comes with choice, context, and control.” — Anonymous

The Inconclusive: What Are We Really Being Offered?

This is where I become less of a scientist and more of a citizen.

Imagine you are a young woman living in rural Malawi. You have heard about this free injection that is said to be effective. However, you recall a time when ARVs were scarce. You have witnessed how stigma can silence the truth.

Do you take the shot?

As a policymaker in Nigeria, you are balancing donor expectations, national priorities, and community concerns. You understand that lenacapavir could save lives, but you also recognize that dependency on it can be dangerous.

Do you endorse the rollout?

These are not hypothetical questions. They are real, urgent, and unresolved.

The Mirror: What Should We Ask Ourselves?

As Africans, as scientists, as educators, we must ask:

1. Are we being seen as partners, or as problems to be managed?

2. Can we trust a system that offers medicine but denies mobility, collaboration, and respect?

3. Will we build health futures rooted in sovereignty, or in dependency?

I want my students to grow up in a world where science empowers, not controls. Where prevention is a choice, not a prescription. Where Africa is not just a testing ground, but a source of innovation, leadership, and dignity.

Conclusion: Between the Lines of Progress

Lenacapavir is a marvel of molecular engineering. But its true test lies not in the lab, but in the lived realities of sub-Saharan Africans. This is a moment to celebrate science, yes — but also to interrogate power, equity, and the politics of health.

I wrote this not to dismiss the drug, but to deepen the conversation. Let this article be a prompt, not a conclusion. Let every reader ask:

What does prevention mean in a world that restricts movement, silences voices, and stereotypes bodies? What kind of future are we building, one injection at a time?

I’ve simplified the science of Lenacapavir in this article to make the message clear. If you enjoyed the scientific aspects, please leave a comment so I can research further and write a more in-depth exploration of Lenacapavir’s science.

Edit: I have written the in-depth research article for Lenacapavir for those who are interested in the nitty gritties of the drug.

A New HIV Shot Arrives in Africa — Is It A Gift or Something Else? was originally published in Microbial Instincts on Medium, where people are continuing the conversation by highlighting and responding to this story.

The problem is that Selection Analysis isn’t always straightforward

In viral genomics, understanding how pathogens evolve under immune pressure or in response to treatment is critical.

One of the most powerful tools for this is codon-level selection analysis, which compares rates of nonsynonymous (dN) and synonymous (dS) mutations to detect signs of adaptive evolution.

But here’s the catch: most tools in this space are either too rigid, too opaque, or too fragile when faced with real-world data. Frame shifts, ambiguous bases and inconsistent annotations break pipelines or skew results.

And even when they work, visualising the results in a biologically meaningful way often requires manual post-processing.

Enter SIMBA (Selection Inference and Mapping for Biological Analysis)

SIMBA is a lightweight, codon-aware Python package designed to make selection analysis both robust and accessible.

It takes in codon-aligned multi-sequence FASTA files and domain annotations, performs population-level dN/dS estimation, flags evolutionary hotspots, and generates publication-ready plots. All from a few lines of code.

from simba.pipeline import run_selection_analysis

run_selection_analysis(
    fasta_path="data/viral_sequences.fasta",
    annotation_path="data/spike_annotations.csv",
    output_dir="results/"
)

That’s it. No GUI. No manual alignment tweaks. Just clean, reproducible science.

Under the hood: Why SIMBA works

SIMBA’s core logic is built around codon-aware aggregation and domain-level mapping. Here’s what it does:

1. Per-codon dN/dS estimation: Calculates observed nonsynonymous and synonymous changes across aligned sequences, normalised by site counts using the Nei and Gojobori, 1986 algorithm
2. Hotspot detection: Flags codons with elevated dN/dS ratios using a threshold (default ≥1.0), indicating potential positive selection.
3. Domain overlay: Maps codons to annotated functional regions (e.g., receptor binding domains, fusion peptides) for biological context.
4. Statistical testing: Applies Kruskal-Wallis tests to compare selection pressures across domains.

Advantages

1. Simplicity: One function call runs the entire pipeline.
2. Robustness: Handles ambiguous bases, frame shifts, and partial codons gracefully.
3. Visualisation: Generates annotated plots with domain overlays and hotspot markers.
4. Reproducibility: CLI support and GitHub packaging make it easy to share and rerun analyses.

Limitations

1. No phylogenetic modelling: SIMBA doesn’t infer trees or account for lineage structure.
2. Assumes clean alignments: While it tolerates noise, it doesn’t perform alignment itself.
3. Threshold-based hotspot detection: May miss subtle signals without model-based refinement.

What you get

1. A mutation inventory .csv table with flagged hotspots and domain annotations.
2. Codon-level plots showing dN/dS ratios, site counts, and domain tracks.
3. Domain-wise boxplots with statistical comparisons.

All outputs are static, publication-ready PNGs — no need to wrangle matplotlib yourself.

Getting Started

SIMBA is open-source and installable via pip:

pip install git+https://github.com/MiltonSimba/SIMBA.git

Then run:

from simba.pipeline import run_pipeline

run_pipeline(
    'fasta filepath',
    '.csv file with annotated genome positions',
    threshold=1.5   # 1.5 is default, user can choose their preference
)

Why it matters

In a world of rapidly evolving pathogens and variants, tools like SIMBA assist researchers in quickly identifying regions under pressure, guiding vaccine design, therapeutic targeting, and enhancing evolutionary understanding.

It’s not a replacement for deep phylogenetic modelling, but it’s a powerful first-pass filter — especially when paired with domain knowledge and structural data.

In conclusion

SIMBA was built with one goal, to make codon-aware selection analysis accessible, robust, and biologically meaningful.

Whether you’re a virologist, bioinformatician, or data scientist exploring evolutionary dynamics, SIMBA gives you a clean starting point — no black boxes, no manual hacks.

And sometimes, that’s exactly what science needs.

References

1. Nei M & Gojobori T (1986) Simple methods for estimating the numbers of synonymous and nonsynonymous nucleotide substitutions. Molecular Biology and Evolution 3:418–426.

SIMBA can be accessed on my GitHub