GenoSiS leverages machine learning–based similarity search to dynamically find patient-matched cohorts across different populations quickly, which could serve as reference cohorts to improve a range of clinical analyses, including disease risk score calculations and dosage decisions.
Abstract
Many patients do not experience optimal benefits from medical advances because clinical research does not adequately represent them. Although the diversity of biomedical research cohorts is improving, ensuring that individual patients are adequately represented remains challenging. We propose a new approach, GenoSiS, which leverages machine learning–based similarity search to dynamically find patient-matched cohorts across different populations quickly. These cohorts could serve as reference cohorts to improve a range of clinical analyses, including disease risk score calculations and dosage decisions. Although GenoSiS focuses on finding genetic similarity within a biobank, our similarity search architecture can be extended to represent other medically relevant patient characteristics and search other biobanks.
The shift toward precision medicine has transformed clinical trial design and drug delivery, yet persistent gaps in data diversity continue to contribute to health disparities. Broad representation and precision stratification are not competing goals; rather, representative datasets are the necessary foundation for ide...
Meghan Davis, Carla Reimold, Kirk Taylor et al.· Biopreservation and Biobanki...· 0 citations
Generation Victoria (GenV), Australia's largest population-representative cohort study, has established a scalable and pragmatic approach to biobanking. This article describes the guiding principles used to overcome previous obstacles and build this unique population-representative mega-cohort birth biobank. By prospec...
K. Lange, Kim L. Powell, Y. Mangwiro et al.· Medical Journal of Australia· 0 citations
This study presents a comprehensive benchmark analysis of InheriNext®, a domain-specific, AI-powered tool designed for phenotype-driven pathogenic variant prioritization. For this study, 7,244 whole exome test cases were generated using phenotype and genotype data from Phenopackets, along with pools of variants from he...
Justin Chang, Kuan Li, Yu-Shen Tsai et al.· Journal of Advances in Artif...· 0 citations
Biological foundation models are new artificial intelligence tools that can learn useful patterns from large biological datasets and apply them to different research questions to help researchers understand disease mechanisms, identify important genetic changes, predict drug responses, and discover treatment targets.
It is shown that SPLENDID significantly improved prediction accuracy over existing methods, particularly for non-European and admixed ancestries, particularly for non-European and admixed ancestries.
Tony Chen, Hao-Yu Zhang, R. Mazumder et al.· Nature Methods· 0 citations
Population heterogeneity in disease biology, drug exposure and treatment response is often addressed only after development has progressed, limiting generalizability and increasing late-stage risk. We propose an anticipate-verify-influence framework to integrate clinically relevant variability from research through dev...
T. Lavé, Bruno Jolain, Ken Wang et al.· Drug Discovery Today· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.