Automated Medical Articles Analysis for Human Genetic Diagnostic

Abstract

DNA can provide vast amount of information about an organism. Geneticists with the use of Next-Generation Sequencing technology are able to analyze genes variations in search of pathogenic mutations in humans. The amount of data obtained from NGS can be huge, what makes it impossible to analyze without automatic support. A number of applications was created to assist NGS data study, but the results of their computations is a list of potential causal variants, which can be treated as a suggestion. For final diagnosis doctors’ experience and broad scientific knowledge is required.To support the process of identifying pathogenic variants we are developing a solution to extract valuable information for selected genetic variants from medical articles. In our article, we would like to present our work on extending existing variant prioritization methods of Exomiser and OmimExplorer applications with the knowledge extracted from articles available in PubMed database using semantic text analysis. The results were compared with Google Scholar search results. We evaluated our work with the data of nine anonymous patients. Such approach enables medical personnel to profit from the newest scientific knowledge during the diagnosis process.