Jingjing Zhai, Aaron Gokaslan, Sheng-Kai Hsu, Szu-Ping Chen, Zong-Yan Liu, Edgar Marroquin, Eric Czech, Betsy Cannon, Ana Berthel, M Cinta Romay, Matt Pennell, Volodymyr Kuleshov, Edward S Buckler
Introduced PlantCAD2, an extended-context, plant-specific DNA language model pre-trained on 65 angiosperm genomes Showed that PlantCAD2 efficiently captures evolutionary conservation, surpassing the 7-billion-parameter Evo2 in 10 of 12 tasks Demonstrated that its 8,192-bp context window improves accessible chromatin prediction in large genomes such as maize
Flowering plants (angiosperms) exhibit extraordinary species diversity, ∼200-fold variation in genome size, and relatively compact coding regions, presenting both a unique challenge and opportunity for DNA language models. Here, we introduce PlantCAD2, an extended-context, plant-specific DNA language model with single-nucleotide resolution, pre-trained on 65 angiosperm genomes, together with a series of public benchmarks for evaluation. Comprehensive zero-shot testing shows that PlantCAD2 (676 million parameters) efficiently captures evolutionary conservation, surpassing the 7-billion-parameter Evo2 in 10 of 12 tasks. With parameter-efficient fine-tuning, PlantCAD2 outperforms the 1-billion-parameter AgroNT across seven cross-species tasks including chromatin accessible region, gene expression, and protein translation. Its 8,192-bp context window substantially improves accessible chromatin prediction in large genomes such as maize (area under the precision-recall curve [AUPRC] increasing from 0.587 to 0.711), underscoring the importance of long-range context for modeling distal regulation. These results establish PlantCAD2 as a powerful and versatile foundation model for plant genome annotation and interpretation across diverse species.