Just as a socially cohesive community with active interaction among its members can foster healthier individuals, could the cohesion of a genetic community also influence a person’s health?
A research team led by Professor Lee Do Heon of the Department of Bio and Brain Engineering at KAIST announced their findings that by identifying less cohesive genetic communities within an individual’s genetic network, they can predict drugs suitable for the patient with four times more accuracy.
The researchers anticipate that this technology could advance personalized medical technology.
With the aging population and changes in lifestyle habits, the incidence of many complex diseases, such as cancer, cardiovascular diseases, metabolic diseases, etc., is significantly increasing. To enhance the treatment effect, researchers put much effort into personalized medicine tailored to individual patient characteristics.
Lee’s research team developed a technology named COSINET (Community Cohesion scores in Individualized Gene Network Estimated from Single Transcriptomics Data), which can accurately measure the cohesion of each genetic community in the intricately constructed personalized gene network.
The research team built a gene network of normal tissues based on significant correlations in gene interactions, using hundreds of normal tissue gene expression data. They modeled the correlation shown in the gene interactions of the genetic communities through linear regression analysis and statistically analyzed whether the gene expression of individual patients follows this predictive model. Through this, they built a personalized gene network by removing gene pairs, whose interaction is lost explicitly in patients, from the normal tissue gene network.
Furthermore, they accurately measured the impact of lost gene interactions on the weakening of the cohesion of each genetic community in the personalized gene network based on the shortest distance between genes.
The researchers demonstrated that they could explain patient-specific disease mechanisms by identifying genetic communities with significantly reduced cohesion for individual patients. By finding genes that significantly contribute to weakening cohesion in these genetic communities, they developed a patient-specific drug target discovery technology that is about four times more effective than existing technologies.
Lee said, “Currently, single-gene-based biomarkers used for personalized medicine in clinical practice have limitations in fully capturing the heterogeneity and complexity of complex diseases.” He added, “Complex diseases involving multiple genes should be viewed from a systemic perspective considering interactions between genes rather than individual genes. Therefore, this study, which measures the cohesion of genetic communities in personalized gene networks, could open a new perspective for realizing personalized medicine for complex diseases.”
This research, jointly conducted by Lee and PhD candidate Wang Seung Hyun, was published in the May 2024 issue of Briefings in Bioinformatics, a top academic journal in the field of bioinformatics published by the University of Oxford, and was announced online on April 15.
Most Commented