The ability to recognize and avoid pathogens is essential for survival. In some cases, animals recognize molecules produced by pathogens, allowing them to respond immediately. In other cases, animals learn to avoid the pathogen after being exposed to it – a phenomenon called “learned avoidance.”
worm Caenorhabditis elegans can learn to avoid pseudomonas aeruginosaA pathogenic bacterium that causes disease in various species (Zhang et al., 2005). In 2019, Colleen Murphy of Princeton University and colleagues observed that the PA14 strain learned to survive P. aeruginosa Can be transmitted up to four generations C. elegans without new exposure to PA14 (Moore et al., 2019; Kaletsky et al., 2020). This transgenerational epigenetic inheritance allows animals that have never encountered PA14 to benefit from the experiences of previous generations.
Recently, Craig Hunter and colleagues at Harvard University questioned the legacy of learned avoidance in the F2 generation (Ganey et al., 2025). While they learned abstinence from PA14 in the parents (P0) and their progeny (F1), they did not observe abstinence in the F2 generation. The Murphy group responded, arguing that the inability to observe transgenerational epigenetic inheritance was due to changes made by the Hunter group to the original experimental protocol (Kaletsky et al., 2025). Now, in eLife, Andres Vidal-Gedia and colleagues at Illinois State University – Alima Akinosho, Joseph Alexander, and Kyle Floyd – report that they have confirmed the Murphy group’s findings that avoidance learned from PA14 is passed on to the F2 generation (Akinosho et al., 2025).
In a standard avoidance assay, test spots of bacteria are placed on opposite ends of an agar plate, and sodium azide – a chemical that immobilizes insects – is added to each test spot (Figure 1). Insects are placed in the center of the plate and allowed to move around, and the number of insects near each test location is scored after one hour (Moore et al., 2019).
Transgenerational inheritance of learned pathogen avoidance.
An avoidance assay (top row) measures the ability to C. elegans Insects to avoid specific bacteria. Worms are first grown on OP50 (orange; top left), a non-pathogenic strain of the bacteria e coli This is the standard laboratory diet of C. elegans, Before transferring them to the test plate. Each bacterial patch is placed at opposite end of a petri dish (top right), the insects are allowed to move around, and their position is recorded after a set period of time. When worms grown on OP50 encounter the pathogenic PA14 strain P. aeruginosa (Green) At first, they are initially attracted to it (top right). However, when worms are exposed to PA14 before the avoidance assay, they avoid PA14 during the assay (second row). This learned defense can be passed on to the next generation (F1; third row) and the generation after that (F2; fourth row), without these insects needing to encounter PA14.
To test learned avoidance, insects are exposed to PA14 for 24 hours, a process sometimes called “training”, after which they are collected and washed. A subset of these insects are transferred to a test plate and allowed to choose between the PA14 and OP50 strains. e coli (which is the standard laboratory diet C. elegansThis is the ancestral generation, and the insects in it will survive PA14.
Eggs (F1) are collected from the remaining insects and transferred to standard OP50 plates. Once they reach the adult stage, the same process is carried out; A subgroup of animals are tested, and eggs are collected from the remaining insects for the next generation (F2). Importantly, after the ancestral generation, worms are no longer exposed to PA14 before testing, but they retain the learned avoidance of the ancestral generation. To measure learned avoidance, all groups are compared to animals whose precursors have never encountered PA14.
Both the Murphy and Vidal-Gadea groups used sodium azide in their trials to immobilize the insects, while the Hunter group immobilized them by lowering the temperature to 4 °C at the end of the trial. With azide, insects that reach test sites before the end of the assay are immobilized. In contrast, when the temperature change method is used, insects may come into contact with test spots and then move away before the end of the assay. The Murphy group proposed that the use of azide ensured that animals were captured in their initial response, preventing them from learning to avoid PA14 after encountering it during the assay. However, they argued, when the temperature-change method is used, insects can encounter PA14, learn from this encounter, and avoid PA14 spots (Kaletsky et al., 2025). Simply put, an encounter with PA14 may inadvertently introduce another source of assay escape. The most significant impact will be on the negative control, where it is assumed that the response of the worms will reflect the response of animals that have never encountered PA14.
Insects that have not previously encountered PA14 are initially attracted to it (Zhang et al., 2005). While the Murphy group consistently observed this attraction in their testing, the Hunter group generally did not (Kaletsky et al., 2025). The Vidal-Gedea group also observed that insects that had not been exposed to PA14 were initially attracted to it, suggesting that it is an important part of the puzzle (Akinoshō et al., 2025). Indeed, when tested directly, the Murphy group did not observe attraction using the temperature-change method (Kaletsky et al., 2025). However, whether the omission of azide alone explains the discrepancy between studies is unclear. In a handful of trials, the Hunter group used azide, but failed to see initial attraction to PA14, or learned avoidance in the F2 generation.
p11, a small RNA produced by PA14, is necessary and sufficient to induce transgenerational epigenetic inheritance (Kaletsky et al., 2020). Furthermore, loss of P11 reduces chemoattraction for PA14 by reducing ammonia production (Maroghi et al., 2024). The Murphy lab proposed that sub-optimal P11 expression could explain the difficulties in reproducing transgenerational epigenetic inheritance and observing initial attraction to PA14. Although they did not measure P11 levels, the Hunter group argued that their ability to observe learned avoidance in the F1 generation ruled out the possibility that P11 expression was insufficient. PA14 growth conditions can alter P11 expression (Kaletsky et al., 2025), but whether low levels of P11 can allow F1 inheritance but not F2 is unknown.
Behavioral tests are extremely subtle, not because they measure effects that are not strong, but because they measure effects that are not strong. C. elegans Are highly attuned to their environment; They integrate myriad environmental signals into an appropriate response. Hunter and colleagues tested possible sources of variability in the laboratory environment, including sources of bacterial and worm strains, but none of these explained the discrepancy in their findings. The sources of variability identified by the Murphy lab (Kaletsky et al., 2025) provide a logical explanation for the differences between the two studies, but other unknown environmental or procedural differences may also have contributed. The results of the Vidal-Gedea group, therefore, represent an important validation of the work of the Murphy group, and support the idea that procedural modifications made by the Hunter group contributed to their inability to observe transgenerational epigenetic inheritance.