Physiological Network From Anthropometric and Blood Test Biomarkers

Area of Science:

• Physiology
• Systems Biology
• Network Science

Background:

• Current physiological research often employs reductionist strategies, focusing on molecular mechanisms.
• Understanding the integrated functioning of physiological variables at the organism level remains challenging.
• Interactions between diverse physiological components are crucial for a holistic systems view.

Purpose of the Study:

• To develop a systems biology approach for representing physiology as an integrated network.
• To build robust physiological networks from biomarker data using Complex Inference Networks.
• To identify network features representative of physiological health.

Main Methods:

• Utilized two independent databases to generate Spearman correlation matrices for 81 and 54 physiological variables.
• Constructed physiological networks by applying a p-value threshold to identify statistically significant correlations.
• Employed unsupervised community detection algorithms to identify functional clusters within the networks.

Main Results:

• Network topology was sensitive to the p-value threshold, but an optimal threshold was identified using stability and connectedness criteria.
• Functional clusters identified through community detection correlated well with established medical knowledge.
• Physiological networks exhibited a topology between random and ordered structures, suggesting robustness and adaptability.

Conclusions:

• Complex Inference Networks provide a robust and visually understandable systems biology framework for physiological data.
• Physiological networks enable the quantification of concepts like homeostasis, aiding in health and disease determination.
• This approach facilitates the exploration of modular functional clusters within physiological systems.