CrossDNA: V1.1.0

Large language models are transforming functional interpretation of genomic sequences. Most DNA sequence language models either treat DNA as unidirectional text or enforce reverse-complement symmetry through data augmentation or equivariance, yet both approximate double-strand interactions implicitly and statically, limiting capture of context-dependent cross-strand information exchange during representation learning. In biological systems, such exchange is dynamically regulated by physical coupling, functional synergy and information transfer, and is fundamental to genomic function. Here we present CrossDNA, an explicit and dynamic language model for DNA cross-strand modelling. CrossDNA uses dual branches with alternating exposure to forward and reverse-complement sequence segments as a duplex-inspired training strategy, and establishes inter-strand communication through a lightweight cross-strand communication module. To capture long-range dependencies, the model combines a recurrent long-context backbone with sliding-window attention, while reverse-complement equivalence, self-distillation and branch-teacher consistency constraints stabilize strand-specific contextual semantics. Across classification, regression and representation tasks, CrossDNA delivers consistent performance gains and improves robustness to sequence orientation, particularly in enhancer prediction, where it identifies features with clear biological significance. Despite having only a few million parameters, CrossDNA matches or surpasses models with hundreds of millions of parameters, reducing training and inference costs. These results advance DNA representation from implicit, static approximation to explicit, dynamic modelling, and provide a foundation for interpreting regulatory logic and prioritizing disease-associated non-coding variants.