Blood-Based Biomarkers Show Promise for Psychosis Risk Prediction

Hallucinations, delusions, and disorganized thinking often evolve gradually, hindering timely identification. In Singapore, 1 in 43 adults has received a diagnosis of psychosis, underscoring the need for objective risk stratification tools. A new study identifies blood-based protein signatures that may forecast transition to psychosis among Asian youths.

Nanyang Technological University (NTU) Singapore, in collaboration with NHG Health’s Institute of Mental Health (IMH) in Singapore, has identified blood-based proteomic biomarkers in plasma that reflect underlying biological processes and may indicate disease risk. Investigators profiled circulating proteins with mass spectrometry-based proteomic analysis and then applied machine learning to develop predictive models. The approach aims to complement symptom-based clinical assessments with quantifiable molecular readouts.

The work drew on IMH’s Longitudinal Youth at Risk Study (LYRIKS), initiated in 2008 to characterize social, clinical, and biological risk factors in youths at ultra-high risk of psychosis. LYRIKS followed 173 participants aged 14–29; 65 were classified as ultra-high risk based on early warning signs, and 13 developed psychosis over two years. Researchers built five prediction models: two using protein panels reported previously in Caucasian cohorts and three derived from the LYRIKS dataset via machine learning and statistical selection.

Models based on Caucasian protein patterns achieved 75% to 81% accuracy on the LYRIKS dataset, , whereas LYRIKS-derived models reached up to 96% accuracy. Although the specific proteins differed between cohorts, similar biological processes—including immune function—were implicated across populations. The findings indicate that population-specific models may improve risk prediction accuracy in Asian groups.

The study was published in Translational Psychiatry. The team noted that larger, independent validations across different populations are needed before clinical implementation. They also acknowledged that the analysis focused on proteins present in most samples, potentially missing lower-abundance yet relevant proteins; future work may broaden analyte coverage and integrate multimodal data, such as genomics, metabolites, and social factors, with explainable artificial intelligence (AI).

"The findings were encouraging as they showed that protein signatures identified in Caucasian populations could be applied more broadly, while also providing insight into how an Asian-specific signature looks and performs. This moves the field closer to molecular-based profiling that can support existing clinical approaches, bringing us a step closer to more personalized, biology-based care," said Assistant Professor Wilson Goh, Lee Kong Chian School of Medicine and NTU Center of AI in Medicine.

“As Singapore’s only national specialist psychiatric institution, IMH is committed to advancing mental health care through research. Research like this helps deepen understanding of complex psychiatric conditions, uncover their biological basis, and support the development of more effective diagnostic tools. This will ultimately improve recovery outcomes for patients,” said Jimmy Lee, Group Chief Research and Innovation Officer, NHG Health, and Senior Consultant and Clinician-Scientist, Institute of Mental Health (IMH).

Related Links
Nanyang Technological University