#artificial intelligence
Sep 2025
Incorporating LLM Embeddings for Variation Across the Human Genome
This work presents one of the first systematic frameworks to generate genetic variant-level embeddings across the entire human genome, using curated annotations from FAVOR, ClinVar, and the GWAS Catalog to construct functional text descriptions for 8.9 billion possible variants.
Hongqian Niu, Jordan G. Bryan, Jacob Williams et al.
· arXiv.org · 1 citation