Mendelian randomization and causal networks for systematic analysis of omics

TL;DR

This paper discusses extending Mendelian randomization and instrumental variable techniques to analyze large-scale omics data, aiming to uncover causal networks and underlying biological mechanisms.

Contribution

It introduces methods to identify and analyze causal networks in large omics datasets, addressing the challenge of complex interconnectivity and unknown structures.

Findings

Extended instrumental variable techniques for large-scale omics

Framework for causal network identification and analysis

Insights into biological mechanisms linking omics and diseases

Abstract

Mendelian randomization implemented through instrumental variable analysis is frequently discussed in causality and recently the number of applications on real data is increasing. However, there are very few discussions to address modern biomedical questions such as the integration of large scale omics in causality. While in the age of large omics, we face several hundred or thousands of components with little knowledge about the underlying structures, the focus of the field is on small scales and mostly with known structures. The availability of large omic data accentuates the need for techniques to identify interconnectivity among the omic components and reveal the principles that govern the relationships. This study extends instrumental variable techniques to identify causal networks in large scales and assess the assumptions. Large-scale causal networks are complex and further analyses are required to uncover mechanisms by which the components are related within and between omics and linked to disease endpoints. This study will review these utilities of causal networks for mechanistic understanding.