Integrating frequency and phenotypic association data to improve interpretation of rare genetic variants

/ABSTRACT At present, the diagnoses of many rare disease patients remain unsolved and the effects of rare variants in common diseases remains unclear. Further, the phenotypic effects of most high-impact mutations are unknown. Accurate methods to interpret genetic variants would enhance many aspects of biomedical research, including genome-wide association studies and functional studies, as well as clinical diagnostics of rare disease patients. In my laboratory, we build tools, methods, and resources to interpret genetic variation that are applicable in clinical practice and human disease research. Beginning with predicted loss-of-function variants, we have built a widely used annotation tool, LOFTEE, that is highly specific in identifying variants triggering nonsense- mediated decay. I have aggregated massive datasets of human genomes and exomes (gnomAD) to build the largest variant frequency maps released to the public, for clinical laboratories and researchers interpreting disease variation. Finally, I have performed massive-scale common and rare variant association analysis on thousands of phenotypes using mixed models to describe the most robust estimates of variant-phenotype associations, powering gene discovery for common diseases. Using state-of-the-art computational methods, this research program will improve the interpretation of variants in the human genome, improving current annotation methods for predicted loss-of-function, missense, and non-coding mutations, incorporating information from genetic reference data as well as biobanks with genome and disease/phenotype data. My research program will build a framework for assessing function for many classes of deleterious variants by integrating frequency data from diverse populations into deep learning frameworks. Finally, this project will integrate rare variant association data into deep learning models to identify loss-of-function-like missense variants. In this way, my research program will improve interpretation of genetic variants found in patients and large cohorts and biobanks, improving clinical genetic practice and human disease research. Project Number: 1R35GM157035-01 | Fiscal Year: 2025 | NIH Institute/Center: National Institute of General Medical Sciences (NIGMS) | Principal Investigator: Konrad Karczewski | Institution: MASSACHUSETTS GENERAL HOSPITAL, BOSTON, MA | Award Amount: $412,500 | Activity Code: R35 | Study Section: Special Emphasis Panel[ZRG1 MGG-F (55)] View on NIH RePORTER: https://reporter.nih.gov/project-details/11020376