Back

Seed microbiota legacy mitigates the effect of drought in wheat

Azarbad, H.; Pande, P. M.; Giard-Laliberte, C.; Agoussar, A.; Dozois, J. A.; L Esperance, E.; Bainard, L. D.; Tremblay, J.; Yergeau, E.

2026-01-16 microbiology
10.64898/2026.01.16.699829 bioRxiv
Show abstract

Seeds carry epiphytes and endophytes microbial partners that can shape plant fitness. However, whether these microbial communities serve as transgenerational memory systems of parental stress remains largely unstudied. Here, we combined multi-year field rainfall manipulation experiments in eastern Canada with a greenhouse experiment in western Canada to test whether seed-associated microbiota transmit drought legacies across plant generations. In the field experiment, reduced rainfall initially decreased yield in the drought-sensitive wheat cultivar (AC Nass), but selected for distinct seed bacterial endophyte communities. In subsequent generations, plants whose seed microbiota retained compositional similarity to these drought-adapted communities showed enhanced yield stability under upcoming water stress. A transgenerational field test confirmed that daughter plants derived from drought-exposed parent plants maintained performance under water limitation, whereas those from wetter origins did not. In an independent greenhouse assay using seeds from Saskatchewan fields differing in long- and short-term irrigation history, AC Nass plants from water-stress legacy sites exhibited higher photosynthetic efficiency, water-use efficiency, and root bacterial diversity under drought. Together, these findings demonstrate that seed-associated microbiota act as ecological archives of stress history, transmitting drought legacies across generations in a cultivar and year-dependent manner.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.