Development and validation of a computable knowledge framework for genomic medicine

点击翻译按钮获取中文摘要

Project Summary/Abstract Genomic medicine is the discipline of interpreting genomic information about an individual as part of their clinical care, for diagnosis, prognosis, or therapeutic decision-making. Integral to the practice of genome interpretation is the collection of multiple lines of evidence from knowledgebases to support or refute the clinical significance of evaluated variants. Modern clinical variant knowledgebases maintain literature and variant coverage that is mostly non-overlapping. This diversity of content causes a known problem in genome interpretation: analysts tasked with assembling a clinical variant report choose to spend considerable time navigating multiple resources and collating evidence, or risk missing critical information by selectively evaluating fewer resources. The resulting effort needed for an analyst to clinically interpret a variant list is known as the interpretation bottleneck, for its rate-limiting role in the clinical evaluation of patient genomes. Data integrators from public and private genomic medicine organizations work to alleviate this bottleneck by developing integrative clinical interpretation applications for use by genome analysts. As new knowledgebases are created, each of these public and private data integrators is left with the task of designing and maintaining another interface for each new resource, leading to combinatorial growth of data harmonization effort across the entire system. This approach is not scalable. The parent R35 is enabling a transition to a scalable, interoperable, and federated genomic data ecosystem from the data integrators and knowledgebases already in existence today through development and validation of a computable knowledge framework for genomic medicine. This objective is being carried out through coordination of research activities with the Variant Interpretation for Cancer Consortium, ClinGen, and the Global Alliance for Genomics and Health. This administrative supplement extends the activities of the parent R35 by applying the developing genomic knowledge framework to the Genome Aggregation Database (gnomAD) a dataset of great value to clinical decision support systems and AI/ML tools used to support clinical variant interpretation. This is achieved through a new collaboration with the gnomAD team to bring the developments of the framework to the gnomAD dataset. Utility of the gnomAD dataset in applied AI/ML tools for genomic medicine will be demonstrated in an augmented intelligence variant classification system. As a result of this administrative supplement, new AI/ML applications dependent upon Population Frequency Evidence will be made possible without need for data harmonization efforts. This will provide a foundation for scalable, AI-assisted classification of variants in genomic medicine pipelines.