Search

Flowering plants have evolved numerous intraspecific and interspecific prezygotic reproductive barriers to prevent production of unfavourable offspring1. Within a species, self-incompatibility (SI) is a widely utilized mechanism that rejects self-pollen2,3 to avoid inbreeding depression. Interspecific barriers restrain breeding between species and often follow the SI × self-compatible (SC) rule, that is, interspecific pollen is unilaterally incompatible (UI) on SI pistils but unilaterally compatible (UC) on SC pistils1,4–6. The molecular mechanisms underlying SI, UI, SC and UC and their interconnections in the Brassicaceae remain unclear. Here we demonstrate that the SI pollen determinant S-locus cysteine-rich protein/S-locus protein 11 (SCR/SP11)2,3 or a signal from UI pollen binds to the SI female determinant S-locus receptor kinase (SRK)2,3, recruits FERONIA (FER)7–9 and activates FER-mediated reactive oxygen species production in SI stigmas10,11 to reject incompatible pollen. For compatible responses, diverged pollen coat protein B-class12–14 from SC and UC pollen differentially trigger nitric oxide, nitrosate FER to suppress reactive oxygen species in SC stigmas to facilitate pollen growth in an intraspecies-preferential manner, maintaining species integrity. Our results show that SRK and FER integrate mechanisms underlying intraspecific and interspecific barriers and offer paths to achieve distant breeding in Brassicaceae crops.

As one of the most important post-translational modification processes, lysine phosphoglycerylation modifications affect many important biosynthetic processes in the human body. However, traditional experimental methods for the recognization of lysine phosphoglycerylation sites are not only expensive but also time-consuming. Computational techniques may provide an economical and efficient way to predict lysine phosphoglycerylation sites. Therefore, it is extremely necessary and meaningful to study and establish prediction models with high accuracy. In the present study, we propose a BERT-based model, BERT_PLPS, which could predict accurately lysine phosphoglycerylation sites. This model extracts amino acid sequence features with three algorithms: CKSAAP, AAC, and BE. Sample equalization is performed using the ADASYN and KNN algorithms. The data are dimensionalized by the ISOMap algorithm, and the features are encoded into feature sequences by an encoder as the input to a BERT-based prediction model. To learn better the intrinsic biological language of lysine, we replaced the original static mask with a dynamic random mask. Compared to other machine learning or deep learning-based models, BERT_PLPS exhibits up to 99.53% accuracy and outperforms the most advanced model (PLP_FS) with an increase of approximately 0.35% on ACC and approximately 0.93% on MCC.