Researchers Submit Patent Application, "Systems And Methods To Generate Local Ancestry Determinations, Polygenic Risk Scores, And Other Useful Information From Genomic Data", for Approval (USPTO 20240304281).

A patent application has been submitted for a system and method to generate local ancestry determinations and polygenic risk scores from genomic data. The application emphasizes the importance of considering ancestral origin when analyzing genomic data, as the accuracy of risk scores can vary depending on the population dataset used. The inventors propose a methodology that incorporates local ancestry information to generate more precise risk scores for individuals. The system utilizes a library of genomic information from non-admixed individuals with known local ancestry to compare with the genomic data of the target individual and generate local ancestry inference information. This method can be used in medical or health-related applications, such as generating risk scores or predicting the likelihood of certain medical conditions. The inventors have developed a two-stage pipeline that combines a base layer determination of recombination distance with a deep learning-based smoothing module to enhance the accuracy of ancestry deconvolution. The method also includes steps for configuring and updating the subpopulation reference datasets and analyzing genomic data for filtering and removing inaccurate data. [Extracted from the article]