Crystal ball time

Editorial https://doi.org/10.1038/s41477-026-02229-4 Crystal ball time Check for updates The great physicist Niels Bohr is reported to have said that “prediction is very difficult, especially about the future”, but that should not stop us trying to guess what 2026 might bring. O ne of the most frequent questions we get asked at conferences is ‘what do you think are the exciting areas of plant research at the moment?’. Implicit in this question is the assumption that as editors we must have a clear vision of everything that is being investigated at this moment. But it isn’t like that at all. Most often, the thing that interests us most is the last study we read (or at least one within the last ten). Like any scientist, what excites us is the unexpected: the result that comes as such a surprise that it changes our views by raising questions we had never thought to ask before. Nevertheless, by looking over our shoulders at the research we have seen during the past few years, there are a few areas in which we expect, or at least hope, to see exciting advances over the next 12 months. One of the fastest moving areas for several years now has been the development of genome editing. This isn’t restricted to plants of course, but the editing of plant genomes has obvious practical applications in the development of new crop varieties. ‘Traditional’ genome editing, using, for example, Agrobacterium-mediated transgene delivery, has limitations such as low transformation efficiency, but using viral vectors to transiently deliver the editing machinery can overcome such drawbacks. In June 2025 we published a Perspective1 by Abraham Steinberger and Daniel Voytas on this approach, following up on a couple of papers that used different viral delivery strategies to edit the genomes of Arabidopsis thaliana2 and wheat3. nature plants Editing efficiency remains lower than is ideal, but we hope to see improvements, as well as extensions to more plant species, soon. A collection of techniques that are quickly becoming almost routine are single-nucleus and spatial transcriptomics. Being able to see exactly which genes are active in every cell of a tissue is providing a level of detail that is completely lost in bulk analysis4. With many of the technical challenges ironed out, these approaches are now being applied to a huge range of systems and questions. Already in this first issue of 2026, we have published a single-cell spatiotemporal transcriptomic study following the infection of potato plants with the devastating late blight pathogen Phytophthora infestans, which shows exactly how heterogeneous the infection process is. While thinking about techniques, we cannot neglect the explosion in the use of large language models (LLMs) and other artificial intelligence (AI) systems in research. Although most discussions in which we are involved concentrate on negative aspects of generative AIs — to what extent they can be legitimately used when preparing research for publication and how to guard against their illegitimate use in manipulating or faking results — their ability to decode all kinds of data is being increasingly utilized. For evidence of the wide applicability of LLMs to data that do not explicitly have a ‘language’, one need look no further than a study5 published in Nature Plants last October that used an LLM to interpret ecological data and thus classify habitat types and predict the presence of unobserved species in a community. The microbiome both on aerial parts of crops and in their below-ground rhizosphere is an increasingly fertile area of research. The full extent and subtlety of the interconnectedness of these agroecological systems is becoming fully appreciated, for example, in this study6 of how herbivory is countered in maize by the stimulation of root exudates that encourage beneficial soil bacteria that in turn reduce herbivore load. Attempts to re-engineer photosynthesis to be more efficient could be compared to the development of atomic fusion as a viable power source: always showing promise but always at least ten years away from becoming reality. Indeed, our Chief Editor was already writing about its potential almost a quarter of a century ago7. However, the prospects of ‘improving’ photosynthesis in crop plants by the inclusion of carbon concentration mechanisms based on the pyrenoid organelles of green algae and hornworts, the carboxysomes of cyanobacteria, or the specialized cell types of C4 or crassulacean acid metabolism plants, seem particularly hopeful at the moment. One step in this direction is the recent work on the molecular engineering of bicarbonate selectivity in the Chlamydomonas reinhardtii formate/nitrite transporter family of proteins, published in this issue. There are plenty of other topics whose developments we are looking forward to following in 2026. Some are broad, such as the effects of climate change on forests (for example, the study by Lu, R. et al. in this issue) and wetlands8, whereas others are more specific, like the emergence of cyclic nucleotides, and particularly cAMP, as second messengers in plant signalling networks9. Whatever we may predict for the coming year, it is certain to be the unanticipated results that will be the most exciting. Published online: 23 January 2026 References 1. Steinberger, A. R. & Voytas, D. F. Nat. Plants 11, 1241–1251 (2025). 2. Weiss, T. et al. Nat. Plants 11, 967–976 (2025). 3. Qiao, J.-H. et al. Nat. Plants 11, 1252–1259 (2025). 4. Nobori, T. New Phytol. 247, 1098–1116 (2025). 5. Leblanc, C. et al. Nat. Plants 11, 2026–2040 (2025). 6. Hu, L. et al. Nat. Plants 11, 1001–1017 (2025). 7. Surridge, C. Nature 416, 576–578 (2002). 8. Li, J. et al. Nat. Ecol. Evol. 9, 1861–1872 (2025). 9. Chen, H. et al. Nature 640, 1011–1016 (2025). Volume 12 | January 2026 | 1 | 1  (...truncated)