Scientists at Stanford University have used artificial intelligence to design the complete DNA of a working virus, a step that could help researchers understand how phages might be turned into treatments for antibiotic-resistant infections. The study, published in the journal Science, produced a functional virus from an AI-designed genome - a demonstration that the technology can learn enough about a biological system to build something that works in the real world.
The researchers worked with ΦX174, a bacteriophage that infects the bacterium E. coli. It's one of the smallest, simplest and best-studied phages known to science, which makes it a solid test case. But it's a long way from the far more complicated phages that scientists want to develop as medicines.
Some potentially useful therapeutic phages are five to 50 times larger than ΦX174, with double-stranded DNA genomes containing hundreds of genes. They carry sophisticated molecular machinery to recognize bacteria, reproduce inside them and overcome bacterial defenses developed over billions of years. A therapeutic phage also needs to work inside a patient, find the right bacteria, and remain useful as bacteria evolve resistance. Much of that complexity remains a mystery - and that's where AI could help.
Where AI fits in phage research
AI is already helping scientists look deeper into phage biology. Software such as AlphaFold can predict the 3D structure of proteins from their amino acid sequences, which is valuable for phage proteins whose functions are still unknown. A gene that once appeared to be little more than a mysterious stretch of DNA can now hint at the shape of the protein it produces and what that protein might do.
AI doesn't replace experiments, but it provides new hypotheses to test. At the Becky Mayer Centre for Phage Research at the University of Leicester, researchers isolate and study phages that infect E. coli, Klebsiella and Pseudomonas - harmful bacteria that are hard to treat. When studied in detail, these phages keep showing unexpected biology: different viruses recognize different parts of bacterial cells, use different tricks to overcome bacterial defenses, and sometimes behave differently in conditions that resemble the human body rather than a lab.
That diversity gives AI something valuable: a large library of real biological solutions. Instead of just asking AI to design a new phage genome, researchers could use it to connect DNA sequence, protein structure and what the phage actually does. For example, could AI identify the genetic features that determine which bacteria a phage can infect, or reveal which proteins help it defeat bacterial defenses, or explain why some phages work well alongside certain antibiotics? For professionals looking to explore AI's applications in research, resources like AI for Research Scientists cover similar ground.
The ultimate question is what separates a phage that looks promising in a lab experiment from one that actually works in the complicated environment of a living patient. AI might spot patterns in vast biological diversity that humans miss. It could predict which natural phages are most promising, identify overlooked proteins, suggest where to find useful phages and eventually propose changes that make them more effective. But every prediction would still need to be tested in the real world.
A feedback loop between computers and experiments
That back and forth could be the most exciting part. AI suggests a biological rule or makes a prediction, scientists test it, and the results feed back into the next round of analysis. The Stanford study opens an important question: has AI learned general rules about how phage genomes work, or does it simply reproduce a well-understood single case? We don't know yet, and the way to find out is to give AI much more biology to work with.
Phages provide an almost unimaginably rich source of data. Nature has spent billions of years experimenting with different ways for viruses to infect bacteria, reproduce, evade defenses and survive in different environments. Scientists can use AI to make sense of that natural experiment.
Why this matters for science and research professionals
For scientists working outside phage biology, the deeper lesson is about how AI transforms research strength. Large biological datasets - sequence libraries, protein structures, experimental observations - hold patterns that exceed human perception. The phage work suggests that feeding those collections into AI will not just automate analysis but help researchers ask better questions about the systems they study. The goal isn't simply AI-generated designs; it's uncovering the natural rules hidden in the data. For scientists in any field working with data, members of an AI for Scientists course could be useful for learning those methods. But the payoff applies broadly: AI won't remove the need for experiments. It will make the next experiments better.
Your membership also unlocks:
