Unified knowledge-driven network inference from omics data
• This paper introduces CORNETO, a novel framework for knowledge-driven network inference that integrates omics data with prior knowledge through constrained optimization.
• CORNETO offers a unified mathematical approach, supporting multi-sample inference across different network types, including undirected, directed, signed graphs, and hypergraphs.
• The framework bridges the gap between data-driven and knowledge-based methods, addressing the challenges of sparse data and noisy networks by leveraging shared patterns across samples.
• CORNETO improves performance in protein-protein interaction networks, intracellular signaling, and metabolic flux modeling, outperforming traditional single-sample inference methods.
• The multi-sample extensions reduce variability and overfitting, producing more interpretable and biologically consistent networks, as demonstrated through simulated and real-world datasets.
• CORNETO is implemented as an open-source Python package, offering flexibility for customization, reuse, and compatibility with multiple mathematical solvers, enhancing accessibility for the research community.
• Applications include multi-condition Flux Balance Analysis (FBA), metabolic network reconstruction from omics data, and intracellular signaling inference from cancer datasets.
@saezlab @JulioSaezRod @attila_gbr @PabloRMier
💻Code:
github.com/saezlab/corneto
📜Paper:
biorxiv.org/content/10.1101/…
#Bioinformatics #NetworkInference #OmicsData #SystemsBiology #MachineLearning #AIforBiology #NetworkModeling