Abstract
Genomics has developed in step with progress in computing. As computational capabilities have grown, analyses have expanded from simple statistics to artificial intelligence (AI)-based approaches within genomics. The decline in sequencing costs has led to the accumulation of diverse genomic datasets, rapidly accelerating AI for genomic analysis. AI models are now developed and applied across many functional domains, including the prediction of transcription factor binding sites, epigenetic elements, DNA methylation, and noncoding sequence functional annotation. With the maturation of architectures such as deep neural networks, convolutional neural networks, recurrent neural networks, and transformers, many genomic models now accommodate longer inputs, capture long-range context, and integrate complex multi-omics data, thereby steadily improving predictive accuracy. Moreover, the emergence of generative AI has enabled models that can simulate and design genomic sequences. The introduction of generative AI into genomics goes beyond inferring function to the capability of replicating functional genomes. These advances will help advance genome interpretation and accelerate our ability to chart and navigate the genomic landscape.