Recent advancements in biotechnology have introduced several innovative frameworks and models aimed at enhancing genomic research and protein design. One notable development is the Discriminative Network Embedding (DNE), a self-supervised learning framework that improves the analysis of protein-protein interaction networks. This method characterizes nodes through a nonlinear contrast between the representations of direct neighbors and more distant nodes. Additionally, the Evo model, a 7-billion parameter genomic language model, facilitates 'semantic mining' to design novel DNA sequences with specific functions. Evo enables genomic autocomplete, guiding the generation of synthetic genomes and creating the SynGenome database, which contains over 120 billion base pairs of AI-generated DNA sequences. Another key innovation is the Knowledge Graph GWAS (KGWAS), which integrates genetic variant data with functional genomics knowledge to enhance association testing and improve the identification of disease-associated variants. KGWAS constructs a comprehensive knowledge graph that includes 70 variant annotations, 40,546 gene-level annotations, and 11 million interactions across 55 relation types, thereby aggregating diverse biological evidence for better statistical power in genomic studies.