Predicting structure-targeted food bioactive compounds for middle-aged and elderly Asians with myocardial infarction: insights from genetic variations and bioinformatics-integrated deep learning analysis

Abstract

Myocardial infarction (MI) is a significant global health issue. Despite the advances in genome-wide association studies, a complete genetic and molecular understanding of MI is elusive and needs to be fully explored. This study aimed to elucidate the genetic framework of MI and explore the potential health benefits of natural compounds (NCs). The genetic architecture of MI was explored using data from the Korean Genome and Epidemiology Study. We pinpointed crucial protein-coding genes related to MI by multi-marker analysis of genomic annotation for gene-based analysis. The bioinformatics-integrated deep neural analysis of NCs (BioDeepNat), a novel disease discovery application, was utilized to assess the influence of NCs on MI-related target proteins and validated with molecular docking analysis. The BioDeepNat application revealed significant NCs on MI-related target proteins, such as E-resveratrol, epicatechin 3-gallate, and kaempferol. The E3 region of RNF213 protein with a point mutation (Arg4810Lys) had different binding energies with NCs, such as ursolic acid and olean-12-en-28-oic acid, compared to the wild type. However, ginsenosides, eleutheroside, oleanolic acid, and hederagenin showed similar binding energies to wild and mutated types of RNF213 protein. The predicted NCs were primarily sourced from foods such as common grapes and teas. Aromatic hydrocarbons are frequently observed as the prevalent functional groups with high binding affinity for NCs in a molecular docking analysis. In conclusion, the proteins encoded by these genes identified by gene-based analysis interacted with several NCs with health promotion found in day-to-day foods, particularly E-resveratrol and kaempferol. This understanding offers promising directions for precision nutrition strategies in MI.