Computational Biology and Chemistry · Nov 1, 2026

Monotonic Additive Modeling Reveals Candidate Biomarkers of Cardio-Renal-Metabolic Disorder Susceptibility

By Sharma Y, Das A

Cardiovascular disease, chronic kidney disease and type 2 diabetes mellitus form an interconnected disease continuum driven by shared risk factors, overlapping molecular pathways and reciprocal progression. This convergence has led to the recognition of cardio-renal-metabolic disorder (CRMD) as a unified systemic condition in which dysregulation of metabolic, vascular, inflammatory and renal pathways precedes organ-specific pathology. Despite this understanding, current diagnostic strategies remain disease-centric and are typically initiated after substantial tissue damage, limiting opportunities for early intervention. Identification of early non-invasive biomarkers capturing shared CRMD susceptibility is therefore of utmost clinical importance. In this study, we developed an explainable monotonic additive machine learning framework to identify blood-based molecular biomarkers associated with CRMD susceptibility. Transcriptomic profiles representing cardiovascular, renal and metabolic disease states were analysed within a unified CRMD paradigm. Following transcriptomic harmonization and batch correction, a monotonic additive XGBoost model combined with integrated biomarker score, gene-response analysis and differential expression analysis was implemented to identify biologically interpretable and directionally consistent biomarkers reflective of shared CRMD-associated dysregulation. Six biomarkers, including EEF1D, RPL27A, ERO1LB, JUN, POSTN and SERPINE2, were identified as potential CRMD susceptibility biomarkers spanning translational regulation, proteostasis, endoplasmic-reticulum function, systemic inflammation, endothelial dysfunction, vascular dysfunction, extracellular remodeling and fibrosis. Importantly, these biomarkers have potential for blood-based detection using established assays such as ELISA, multiplex immunoassays and quantitative PCR, underscoring their translational potential for early CRMD screening, susceptibility assessment and risk stratification.

doi.org/10.1016/j.compbiolchem.2026.109530PubMed 42976572Biomarkers