Genome foundation models hold great potential for biological research, but modeling diverse complex genomes and generating controllable genome-scale sequences remain challenging. We present DeepD, a phylogeny-informed sparse mixture-of-experts model operating at single-nucleotide resolution with a 1-million-token context and reduced active computation. Pretrained on unselected whole genomes across all domains of life, DeepD learns genic and intergenic features and supports cross-species sequence-to-function prediction. DeepD performs well in protein fitness and disease-variant effect prediction in human, including BRCA1 and BRCA2, with particular strength in promoters and transcription-factor binding sites. Expert routing captures exon–intron boundaries, transposable elements, and distal cis-regulatory elements. DeepD generates experimentally validated maize core promoters with specified strengths. Its Context–Anchor–Body strategy generates complete yeast chromosomes and megabase-scale Arabidopsis segments, preserving gene structure and order while introducing natural-range variation. DeepD thus unifies accurate prediction with controllable genome-scale generation, advancing the predictive and generative design of complex genomes.

Genome foundation models hold great potential for advancing biological research, but modeling diverse complex genomes and generating controllable genome-scale sequences remain challenging. We present DeepD, a phylogeny-informed sparse mixture-of-experts model operating at single-nucleotide resolution over a 1-million-token context while reducing active computation. Pretrained on unselected whole-genome sequences across all domains of life, DeepD learns genic and intergenic sequence features and supports broad sequence-to-function prediction across species. DeepD performs well in human protein fitness and disease variants effect prediction, such as BRCA1 and BRCA2, with particular strength in regulatory regions, such as promoter and transcription-factor binding sites. Expert routing captures exon-intron boundaries, transposable elements, and distal cis-regulatory elements. DeepD also generates experimentally validated core promoters with specified transcriptional strengths. Finally, the Context-Anchor-Body strategy enables generation of complete yeast chromosome sequences and megabase-scale Arabidopsis segments, preserving gene structure and order while introducing variations within natural ranges. Together, DeepD is an efficient genome foundation model that unifies accurate sequence-to-function prediction with controllable genome-scale generation, advancing predictive and generative design of complex genomes.