MR-GGI: accurate inference of gene–gene interactions using Mendelian randomization

Wonseok Oh, Junghyun Jung, Jong Wha J. Joo

Research output: Contribution to journalArticlepeer-review

Abstract

Background: Researchers have long studied the regulatory processes of genes to uncover their functions. Gene regulatory network analysis is one of the popular approaches for understanding these processes, requiring accurate identification of interactions among the genes to establish the gene regulatory network. Advances in genome-wide association studies and expression quantitative trait loci studies have led to a wealth of genomic data, facilitating more accurate inference of gene–gene interactions. However, unknown confounding factors may influence these interactions, making their interpretation complicated. Mendelian randomization (MR) has emerged as a valuable tool for causal inference in genetics, addressing confounding effects by estimating causal relationships using instrumental variables. In this paper, we propose a new statistical method, MR-GGI, for accurately inferring gene–gene interactions using Mendelian randomization. Results: MR-GGI applies one gene as the exposure and another as the outcome, using causal cis-single-nucleotide polymorphisms as instrumental variables in the inverse-variance weighted MR model. Through simulations, we have demonstrated MR-GGI's ability to control type 1 error and maintain statistical power despite confounding effects. MR-GGI performed the best when compared to other methods using the F1 score on the DREAM5 dataset. Additionally, when applied to yeast genomic data, MR-GGI successfully identified six clusters. Through gene ontology analysis, we have confirmed that each cluster in our study performs distinct functional roles by gathering genes with specific functions. Conclusion: These findings demonstrate that MR-GGI accurately inferences gene–gene interactions despite the confounding effects in real biological environments.

Original languageEnglish
Article number192
JournalBMC Bioinformatics
Volume25
Issue number1
DOIs
StatePublished - Dec 2024

Keywords

  • Gene regulatory network
  • Gene–gene interactions
  • Mendelian randomization
  • Yeast GRN

Fingerprint

Dive into the research topics of 'MR-GGI: accurate inference of gene–gene interactions using Mendelian randomization'. Together they form a unique fingerprint.

Cite this