(24-01-25) Successful Master research project defense (MSc. Data Science for Decision Making and MSc. Artificial Intelligence)

Abstract

This work presents a system for automated information extraction from Genome-Wide As- sociation Studies (GWAS) literature, focusing on gene variants and their associated traits. By leveraging a Unified Information Extraction (UIE) model and a domain-tailored schema, our approach systematically identifies key entities (e.g., gene variants and phenotypes) and their relationships, storing them in a searchable knowledge graph. The resulting tool aids researchers in rapidly uncovering and organizing genotype–phenotype insights from GWAS papers, demon- strating both high accuracy and strong coverage in a specialized scientific domain.