CMB_2024v14n2

Computational Molecular Biology 2024, Vol.14, No.2, 84-94 http://bioscipublisher.com/index.php/cmb 84 Review and Progress Open Access Advances in Causal Inference Methods for Biological Network Analysis Jiefu Lin, Kaiwen Liang Hainan Key Laboratory of Crop Molecular Breeding, Sanya, 572025, Hainan, China Corresponding author: kaiwe liang@hitar.org Computational Molecular Biology, 2024, Vol.14, No.2 doi: 10.5376/cmb.2024.14.0010 Received: 20 Feb., 2024 Accepted: 01 Apr., 2024 Published: 21 Apr., 2024 Copyright © 2024 Lin and Liang, This is an open access article published under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Preferred citation for this article: Lin J.F., and Liang K.W., 2024, Advances in causal inference methods for biological network analysis, Computational Molecular Biology, 14(2): 84-94 (doi: 10.5376/cmb.2024.14.0010) Abstract This study summarizes various causal inference methods for biological network analysis, including Bayesian networks, Granger causality, and structural equation modeling (SEM). We explored the application of these methods in integrating multiple omics data and how to overcome the challenges posed by high-dimensional data. Especially, the application of causal inference in disease network analysis demonstrates its potential in identifying key genes, revealing disease mechanisms, and promoting precision medicine. We also evaluated the latest developed computing tools and open-source platforms, which make large-scale data processing more efficient and user-friendly. In the future, the development of causal inference will further rely on the integration of emerging technologies such as machine learning and single-cell omics to promote a deeper understanding of complex disease mechanisms. Keywords Causal inference; Bayesian networks; Granger causality; Structural equation modeling; Gene regulatory networks 1 Introduction Biological networks represent the complex interactions between biological entities, such as genes, proteins, or metabolites, within a system. These networks include gene regulatory networks (GRNs), metabolic networks, and protein-protein interaction networks, all of which are essential for maintaining cellular functions. High-throughput technologies like next-generation sequencing, mass spectrometry, and single-cell RNA sequencing (scRNA-seq) have facilitated the reconstruction of these networks by enabling the collection of large-scale biological data. The analysis of such data helps reveal the structure and function of biological networks, providing critical insights into cellular processes and disease mechanisms. However, the sheer complexity of these networks, their dynamic nature, and their non-linear interactions pose significant challenges to network inference and analysis (Shojaee & Huang, 2023). Accurate modeling of these networks is crucial for understanding biological systems and developing new therapeutic strategies. Causal inference is critical for distinguishing between correlation and true cause-effect relationships in biological networks. Unlike correlation-based methods, which only identify associations, causal inference provides insights into how biological systems respond to perturbations, such as gene knockouts or drug treatments. This is particularly important in understanding disease progression and treatment responses, as well as identifying potential therapeutic targets. Recent advances in causal inference methods, including Granger causality and Bayesian networks, have enhanced our ability to analyze high-dimensional biological data and uncover the underlying causal mechanisms. These methods have been applied successfully to infer gene regulatory interactions, identify key regulators, and predict the outcomes of therapeutic interventions (Ahmed et al., 2020). As a result, causal inference is a valuable tool in systems biology, particularly in precision medicine. This study provides an in-depth analysis of the latest advances in causal inference methods in biological network analysis: exploring the latest algorithmic innovations, including algorithms that address nonlinear interactions and integrate multi omics data; Secondly, investigate the application of these methods in disease research, particularly in identifying therapeutic targets for complex diseases such as cancer and neurodegenerative diseases; Finally, we will discuss the challenges and future directions faced in this rapidly developing field, such as processing high-dimensional, noisy datasets and integrating dynamic, time-resolved data, with the aim of advancing biomedical research.

RkJQdWJsaXNoZXIy MjQ4ODYzNA==