The maze runner: navigating through basic kinetics to AI models of human metabolism pathology

Abstract

Common human metabolic pathologies such as obesity and diabetes are extremely widespread non-infectious diseases. Even though they are thoroughly researched, there are still some unmet clinical needs and both physicians and clinical researchers are desperately looking for answers. Due to recent development of artificial intelligence (AI) technologies and collected big data of clinical cases there is a current trend in using solely advanced AI methods in theoretical modeling, drug development and model parameter estimation. However, not an every aspect and a process involved in metabolic pathologies are covered by a good quality clinical data of a big volume. It poses certain challenges to application of AI methods to such processes and traditional kinetics approaches return to the forefront. It is especially relevant when some structural abnormality of cell topology rather than a gene dysfunction appears to be a possible reason of a condition. In this review, we discuss various approaches to human metabolism disorders modeling and their relevance in different situations.