Omni-DNA: A Unified Genomic Foundation Model for Cross-Modal and Multi-Task Learning

TL;DR

Omni-DNA introduces a unified transformer-based model capable of multi-task and cross-modal genomic analysis, reducing the need for separate fine-tuning and enabling complex applications like DNA-to-text and DNA-to-image mapping.

Contribution

The paper presents Omni-DNA, a novel multi-task, cross-modal genomic foundation model that achieves state-of-the-art results and demonstrates broad applicability across diverse genomic tasks.

Findings

Achieves state-of-the-art performance on 18 out of 26 genomic tasks.

Successfully performs multi-task fine-tuning on 10 epigenetic modification tasks.

Demonstrates cross-modal capabilities with DNA-to-text and DNA-to-image tasks.

Abstract

Large Language Models (LLMs) demonstrate remarkable generalizability across diverse tasks, yet genomic foundation models (GFMs) still require separate finetuning for each downstream application, creating significant overhead as model sizes grow. Moreover, existing GFMs are constrained by rigid output formats, limiting their applicability to various genomic tasks. In this work, we revisit the transformer-based auto-regressive models and introduce Omni-DNA, a family of cross-modal multi-task models ranging from 20 million to 1 billion parameters. Our approach consists of two stages: (i) pretraining on DNA sequences with next token prediction objective, and (ii) expanding the multi-modal task-specific tokens and finetuning for multiple downstream tasks simultaneously. When evaluated on the Nucleotide Transformer and GB benchmarks, Omni-DNA achieves state-of-the-art performance on 18 out of 26 tasks. Through multi-task finetuning, Omni-DNA addresses 10 acetylation and methylation tasks at once, surpassing models trained on each task individually. Finally, we design two complex genomic tasks, DNA2Function and Needle-in-DNA, which map DNA sequences to textual functional descriptions and images, respectively, indicating Omni-DNA's cross-modal capabilities to broaden the scope of genomic applications. All the models are available through https://huggingface.co/collections/zehui127