Only a third of Fusarium oxysporum formae speciales are monophyletic: a per-label and per-deposit audit of 363 public genomes
By auditing 363 public *Fusarium oxysporum* genomes, this study reveals that only one-third of formae speciales are monophyletic and that the current classification system is plagued by significant phylogenetic conflicts, taxonomic redundancy, and mislabeled or non-target genomes, arguing that *forma specialis* should be treated as an operational pathological category requiring independent evidence rather than a phylogenetic one.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer
Fungi that live in the soil are often invisible to the human eye, yet they hold immense power over the food we eat. Among the most notorious of these are members of the Fusarium oxysporum complex, a vast family of microscopic organisms that cause vascular wilt, a disease that clogs the water-conducting tissues of plants and can kill entire crops. For over a century, scientists have sorted these fungi into groups based on which plants they attack. A fungus that destroys tomatoes is placed in one category, while one that kills bananas goes into another. This system, known as the "forma specialis" classification, has been the standard way to organize these pathogens. It is a practical tool for farmers and plant breeders, but it was never intended to be a map of evolutionary history. Because these fungi look identical under a microscope and reproduce mostly without sex, their true family tree has remained difficult to read. The central question has long been whether the groups defined by the plants they attack actually correspond to distinct branches on the tree of life, or if the system is merely a collection of unrelated organisms that happen to share a similar habit.
A recent study by researchers at Ege University in Turkey has taken a massive, systematic look at this question using the genetic blueprints of hundreds of these fungi. The team gathered 816 publicly available genome sequences from a global database and filtered them down to a high-quality set of 363 distinct fungal genomes. They then compared the core genetic material shared by all of them, a set of nearly 1.5 million letters of genetic code, to build a precise family tree. Their goal was to see if the traditional labels—such as "tomato-specialist" or "banana-specialist"—matched up with the actual evolutionary branches. The results were stark and clear. Of the 36 different groups they could test, only 12 were true evolutionary families. In other words, only about one-third of the named groups were monophyletic, meaning they all descended from a single common ancestor. The vast majority were scattered across the tree, appearing in many different places, which indicates that the ability to attack a specific plant has evolved independently many times or has been swapped between unrelated strains.
The researchers found that this confusion was not just a matter of having too few samples. Even when they mathematically reduced the size of the large groups to match the small ones, the pattern of scattering remained. The largest groups, such as those attacking melons, strawberries, and bananas, were broken into as many as eleven separate, independent lineages. This suggests that the ability to infect a specific crop is not a fixed trait of a single family, but rather a flexible skill that can appear in very different genetic backgrounds. Furthermore, the study revealed that the ability to cause disease is carried on a separate, mobile part of the genome that can jump between strains, while the main body of the genome records a different history. This explains why two fungi that look genetically very similar might attack different plants, while two that are genetically distant might attack the same one.
Beyond the evolutionary findings, the audit of the public database uncovered significant issues with the data itself. The researchers discovered that nearly 11 percent of the genomes in their set were exact duplicates of others, meaning they carried no new information. More critically, they identified six genomes labeled as banana-attacking fungi that were actually copies of a famous tomato-attacking strain, likely created through a computer error during data assembly. These six genomes were so similar to the tomato strain that they shared an identical set of disease-causing genes, yet they were being used in research to study banana wilt. Additionally, one genome labeled as a fungus attacking peppers was found to be a completely different organism that did not belong to the Fusarium family at all. None of these errors would have been caught by standard quality checks, as the genetic sequences were complete and well-assembled; the mistakes were purely in the labels attached to them.
The study concludes that the traditional system of naming these fungi by their host plant is a useful operational tool for identifying disease risks, but it is not a reliable guide to their evolutionary relationships. The researchers argue that scientists should treat these labels as hypotheses that need to be verified with genetic evidence, rather than as facts. They have released a cleaned-up, non-redundant set of 324 genomes for the scientific community to use, removing the duplicates and the mislabeled entries. This work highlights a broader lesson for the field of genomics: as the number of available genetic sequences grows, the need for careful auditing of the metadata attached to them becomes just as important as the analysis of the sequences themselves. Without this scrutiny, studies on how these pathogens evolve and spread could be built on a foundation of misidentified data, leading to conclusions that do not reflect biological reality.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.