The global integrative network: integration of signaling and metabolic pathways

Abstract

The crosstalk between signaling and metabolic pathways has been known to play key roles in human diseases and plant biological processes. The integration of signaling and metabolic pathways can provide an essential reference framework for crosstalk analysis. However, current databases use distinct structures to present signaling and metabolic pathways, which leads to the chaos in the integrated networks. Moreover, for the metabolic pathways, the metabolic enzymes and the reactions are disconnected by the current widely accepted layout of edges and nodes, which hinders the topological analysis of the integrated networks. Here, we propose a novel “meta-pathway” structure, which uses the uniformed structure to display the signaling and metabolic pathways, and resolves the difficulty in linking the metabolic enzymes to the reactions topologically. We compiled a comprehensive collection of global integrative networks (GINs) by merging the meta-pathways of 7077 species. We demonstrated the assembly of the signaling and metabolic pathways using the GINs of four species—human, mouse, Arabidopsis, and rice. Almost all of the nodes were assembled into one major network for each of the four species, which provided opportunities for robust crosstalk and topological analysis, and knowledge graph construction.

GINv2.0: a comprehensive topological network integrating molecular interactions from multiple knowledge bases

“…The credibility of the edges is also important for network analysis, since questionable edges will create misleading path when conducting path-related network analysis, as evidenced in our previous work 22 . In the comparative analysis of different databases, repeated edges may be more credible since it has been repeatedly validated by different databases.…”

“…In this work, we compiled a much more comprehensive GIN for human compared with the previous version. The previous GIN 22 for human was built only upon KEGG, which includes 5145 genes and 1501 metabolites. In the present work, we compiled a new GIN for human from ten different databases, which involved 6330 genes and 3579 metabolites, with 23.0% and 138.4% increase, respectively.…”

“…Therefore, the intermediate nodes which come from the intermediate state of the molecules capture the relationships between molecules in real world, and enables both signaling and metabolic reactions to be considered as chemical reactions, facilitating storage in SIF-like format. By converting the pathways into meta-pathways and merging them, we have successfully built GINs for 7077 species based on KEGG 22 .…”

COSMIC-dFBA: A novel multi-scale hybrid framework for bioprocess modeling

“…In the current implementation of COSMIC-dFBA, we circumvent this by defining the cellular objectives for each type of phase shift in advance. However, automated prediction of changes in metabolic task priorities in response to phase shifts will require an overlay of the signaling (Lin et al, 2022; Sompairac et al, 2019) and gene expression networks (Pio et al, 2022) on to existing models of metabolism in the absence of fully descriptive whole-cell models (Ahn-Horst et al, 2022; Karr et al, 2012). Such models will expand the predictive capabilities of COSMIC-dFBA to predict heterogeneity in cell populations in large-scale bioreactors arising from non-homogeneous mixing and poor local oxygen transfer.…”