Demonstrating the potential of untargeted hair proteomics for personalized biomarkers in stress-associated disorders

This study explores the potential of untargeted hair proteomics as a novel biomarker source for stress-associated psychopathology, specifically focusing on non-suicidal self-injury disorder (NSSID) in women. Using machine learning, the researchers analyzed protein profiles from hair segments of 68 participants (36 with NSSID and 32 healthy controls), identifying 611 proteins for analysis. The model achieved an 84.4% accuracy in classifying clinical groups, outperforming other models based on stress-related proteins and simulated hair cortisol. Key predictive proteins were linked to pain perception, oxidative stress, and cholesterol homeostasis. Approximately 15% of proteins showed significant differences between groups, with ribosomal function emerging as a strong candidate related to depression. These findings suggest that hair proteomics could serve as a promising, non-invasive biomarker for psychiatric research, offering potential clinical applications in risk assessment and personalized interventions.