Medical Genomics Unit Analytic Projects

National Human Genome Research Institute

Investigators

Linked publications & trials

Abstract

Following the relatively recent establishment of the Medical Genomics Unit (MGU), our research group this has continued to focus on several related projects, all of which are primarily computational, and which are tied to leveraging underlying methods involving forms of artificial intelligence, especially deep learning. We have used these methods to interrogate a number of data types, all of which relate to aspects of genetic conditions, with the overall goal of better understanding the causes, manifestations, management, and underlying biology of these conditions. These investigations may lead to implementation research. The MGU has led several studies that have been formally approved by the NIH IRB as "exempt" studies, and the MGU has recently garnered approval of a non-exempt IRB-approved research protocol. For projects to date, data types investigated by the MGU include images (eg, picture of skin lesions, optical coherence tomography from ophthalmologic assessment, facial images, etc.), text-based data (eg, standardized terminology for phenotypic terms, clinical notes and descriptions, etc.), and other data types. This work has enabled the MGU to publish several papers on the above topics, including a scoping review on deep learning in medical genetics, as well as papers involving our group's original research findings. We have also established collaborations with other NIH and extramural groups with aligned interests, and anticipate that such work can yield fruitful results in the years to come.