Biological complexity explicitly occurs through non-linear interactions mostly entangled in nature. This complexity comprises many interactions among entities (viz., genes, proteins, metabolites, and species) at various spatial and temporal scales as complex adaptive systems showing characteristic features like self-organisation, modularity, emergence, non-linear interactions, collective response, and adaptation. The theory of complex networks provides an appropriate formal framework for modelling of such complex systems in order to obtain meaningful insights into biological complexity at the local or gene family level (Mallikarjuna et al., 2020; Mallikarjuna et al., 2022) and at the global scale (Sharma et al., 2021). The ocean of biological data generated by high-throughput technologies in the current genomics era have led to the application of various network-theoretic empirical investigations, in which the formal framework is used to obtain meaningful insights into system complexity. Our effort to pool studies on network-theoretic approaches in biology to the understanding of biological complexity has resulted in the compilation of ten research studies in the current Research Topic entitled Application of network theoretic approaches in biology, which are broadly categorised and highlighted under the following headings.

The availability of various user-friendly approaches and software applications has expanded the use of network-theoretic approaches to understand the complex biological process at individual and systems levels. Four articles in the current Research Topic deal with web tools and methods. A user-friendly web tool for the construction of gene regulatory networks, known as the “Consensus Approach for Gene Regulatory Network Construction” (CAGNC), was developed using the R programming language. CAGNC provides a network file with the edge scores representing significant interactions between each gene pair (Sarkar et al.). A new method, PolyReco, was developed to provide a reference model for processing, with automatic labelling of collinear regions and recognition of polyploidy events (Wang et al.). Lee et al. propose a hybrid intelligence platform, StarGazer (https://github.com/AstraZeneca/StarGazer), which provides an interactive dashboard that allows rapid searching for potential novel drug targets and the use of repositioning strategies via the Streamlit tool. Co-expression studies aid in the discovery of network patterns, functional module identification, and trait-linked marker mining at the system level. Finally, Xiao et al. present a knowledge-injected semi-supervised learning (KISL) method (https://github.com/Mowonhoo/KISL.git) for the identification of outstanding modules in a co-expression network. The KISL approach utilises a priori biological information and semi-supervised clustering to solve the issues present in contemporary clustering approaches, such as weighted gene co-expression network analysis (WGCNA).

In conclusion, our Research Topic presents various statistical methods and tools expanding the utility of network-theoretic approaches. Other articles demonstrate the application of various network approaches in developing our understanding of biological phenomena in plant, microbial, and human systems. Nevertheless, for large-scale applications and utilisation of network approaches in biology, there is a further need to undertake some of the following measures, although this is not an exhaustive list. First, more efforts should be made to develop biologist-friendly servers and tools for various types of network analysis, which can allow us to derive meaningful information from the ocean of omics data. Second, there is a need for the development of system-specific network approaches in order to understand species interactions from complex ecological and evolutionary perspectives. Finally, validation of major hubs through genetic and in-depth network-theoretic approaches could demonstrate the biological significance of network-theoretic studies in biology.