Systematic analysis of the dynamic relationship between environment, metabolism, and health condition applying a machine learning approach

Abstract

Since the human genome was sequenced, extensive effort has been placed into mapping the role of genes in the onset of disease. It was expected that the scientific community would be able to explain the cause of disease and understand the genetic basis of health. However, it has found that while the genetic contribution to individual conditions varies, non-genetic factors have far more significant attributable risks, often in the range of 80-90%. The exposome represents an omics-scale characterization of non-genetic contributions to an individual's phenome, including products such as the metabolome. The current project will take an exposome approach, linking individual and societal health to the environment to define fingerprints (molecular signatures) and internal biomarker perturbations. These approaches will require new data analysis strategies, including machine and deep learning artificial intelligence-based methods for large-scale data interrogation. In summary, the present work aims to apply machine learning algorithms to identify specific fingerprints on the internal exposome linked to the external exposome and lifestyle in a group of adults with different lifestyles and residing in the city of Ribeirão Preto, SP, Brazil. (AU)