Employ machine learning methods in combination with bioinformatic analysis of the gene structure of polygenic diseases to develop effective methods for early diagnosis and prognosis of their developmentaper

The study of the features of the genetic structure of cytokines, angiogenesis and vascular remodeling factors in healthy individuals and patients with various forms of diseases - myocardial infarction, breast cancer, rheumatoid arthritis, female infertility of unknown origin and type 2 diabetes mellitus was carried out. 526 healthy and 1074 with various pathology people were examined. The aim of the work was to conduct comparative clinical and genetic studies with the analysis of combined genotypes in complex with various variants of mathematical modeling and machine learning algorithms for the development of new fundamentally methods of early diagnosis and prognosis of the development of polygenic human diseases. The study was conducted using the SNP methods, various variants of data mining and machine learning (simulation modeling, MDR method, bioinformatic analysis of gene networks). It is shown that for the construction of prognostic models of polygenic diseases with the greatest predictive potential, it can be effective to use simulation modeling based on a computational algorithm embedded in the modified procedure of sequential recognition A. Wald. To improve the quality of predictive models, it is necessary to use a combined mathematical modeling method that combines simulation modeling, MDR analysis and graphical visualization of gene networks using the Cytoscape platform. The results of the conducted research can serve as a scientific basis for the development of effective technologies for the prediction, diagnosis and treatment of the polygenic diseases.