Rethinking Genomic Modeling Through Optical Character Recognition

Recent genomic foundation models largely adopt large language model architectures that treat DNA as a one-dimensional token sequence. However, exhaustive sequential reading is structurally misaligned with sparse and discontinuous genomic semantics, leading to wasted computation on low-information background and preventing understanding-driven compression for long contexts. Here, we present OpticalDNA, a vision-based framework that reframes genomic modeling as Optical Character Recognition (OCR)-style document understanding. OpticalDNA renders DNA into structured visual layouts and trains an OCR-capable vision--language model with a \emph{visual DNA encoder} and a \emph{document decoder}, where the encoder produces compact, reconstructible visual tokens for high-fidelity compression. Building on this representation, OpticalDNA defines prompt-conditioned objectives over core genomic primitives—reading, region grounding, subsequence retrieval, and masked span completion—thereby learning layout-aware DNA representations that retain fine-grained genomic information under a reduced effective token budget. Across diverse genomic benchmarks, OpticalDNA consistently outperforms recent baselines; on sequences up to 450k bases, it achieves the best overall performance with nearly $20\times$ fewer effective tokens, and surpasses models with up to $985\times$ more activated parameters while tuning only 256k \emph{trainable} parameters.

DNA sequences can be extremely long, making them difficult for AI models to analyze efficiently. Most existing models process DNA one letter at a time, which can be slow and may struggle to capture useful signals across long distances. We propose OpticalDNA, a method that represents DNA as visual documents so that image-based AI models can process long sequences more efficiently. This representation preserves the original DNA sequence while reorganizing it into a format that supports better access to distant regions. We evaluate OpticalDNA on genomic prediction tasks such as long-range gene regulation and whole-genome phenotype prediction, and find that it achieves strong performance while remaining robust to changes in visual formatting, image resolution, token compression, and cross-species transfer. These results suggest that visual document representations offer a practical new way to model long DNA sequences and may help future AI systems better understand genome-scale biological information.