Demonstrating the Potential of Untargeted Hair Proteomics for Personalized Biomarkers in Stress-Associated Disorders

This study explores the potential of hair proteomics as a biomarker for stress-associated psychopathology, using machine learning to analyze protein profiles from hair segments of 36 women with non-suicidal self-injury disorder and 32 healthy controls. Out of 1114 identified proteins, 611 were analyzed, achieving an 84.4% cross-validated accuracy in classifying clinical groups. Key predictive proteins were linked to pain perception, oxidative stress, and cholesterol homeostasis, with approximately 15% of proteins differing significantly between groups. The findings suggest hair proteomics as a promising non-invasive biomarker source for psychiatric research, highlighting the need for validation in larger cohorts and exploration of clinical applications.