Integrated Chemometric and Machine Learning Analysis Identifies Peripheral Biosignatures Distinguishing Major...

Area of Science:

• Biochemistry
• Psychiatry
• Molecular Biology

Background:

• Major Depressive Disorder (MDD) and Bipolar Disorder (BD) lack objective molecular stratification.
• Clinical overlap between MDD and BD, especially during depressive phases, complicates diagnosis.
• Objective biomarkers are needed for accurate diagnosis and treatment of mood disorders.

Purpose of the Study:

• To explore coordinated peripheral biomarker patterns in MDD and BD using multivariate analysis.
• To identify objective molecular signatures differentiating MDD, BD, and healthy controls (HC).
• To investigate differences in inflammatory indices, trace elements, BDNF, and NLRP3.

Main Methods:

• Cross-sectional study of 151 participants (MDD, BD, HC).
• Profiling of 42 blood-derived parameters including inflammatory indices, trace elements (ICP-MS), BDNF, and NLRP3 (ELISA).
• Analysis using univariate testing, PCA, t-SNE, and PLS-DA with cross-validation and permutation testing.

Main Results:

• Significant inter-group differences (p < 0.05) in 37 of 42 parameters.
• Reduced circulating NLRP3 in both MDD and BD groups compared to HC.
• Elevated inflammatory indices (NLR, SIRI, SII) in MDD; altered zinc and manganese levels in psychiatric cohorts.
• BDNF levels were lower in MDD and higher in BD relative to HC.
• PLS-DA achieved 89.4% classification accuracy for psychiatric disorder vs. controls (AUC-ROC 0.947).

Conclusions:

• Multidomain peripheral biomarker profiling reveals coordinated biochemical differences across mood disorder diagnostic groups.
• These findings suggest the existence of multidimensional peripheral signatures associated with MDD and BD.
• Further research is warranted to validate these exploratory findings for clinical application.