Bioinformatics creates psychosis risk diagnostic

September 23, 2014 − by Suzanne Elvidge − in Big data, Data analytics, Healthcare big data analysis − No Comments

Certain types of psychosis can be prevented or managed, but it is hard to identify who is at risk. A group of US researchers, part of the North American Prodrome Longitudinal Study consortium, has created a blood test that has potential to pick out those who could benefit from prevention interventions, using bioinformatics to tease out patterns from a series of different factors.

Psychosis includes hallucinations or delusions that define the development of severe mental disorders such as schizophrenia. Schizophrenia emerges in late adolescence and early adulthood and affects about one in every 100 people.

The North American Prodrome Longitudinal Study (NAPLS) is a consortium of eight programs looking at the psychosis prodrome, an early set of symptoms that could indicate the onset of schizophrenia. The aim is to understand risk factors and mechanisms for development of psychotic disorders, and the project is funded by the National Institute of Mental Health (NIMH).

In this study, reported in Schizophrenia Bulletin, the teams measured levels of hormones in the blood, and levels of markers for inflammation, immune response, oxidative stress, and metabolism. Using a greedy algorithm (an algorithm that makes the best decision at each step), the researchers analysed a group of 67 people with symptoms that are considered to be indicators of a high risk for psychosis, and selected 15 markers from 117 that could pick out those people who later went on to develop the disorder.

“While further research is required before this blood test could be clinically available, these results provide evidence regarding the fundamental nature of schizophrenia, and point towards novel pathways that could be targets for preventative interventions,” say Diana O. Perkins, MD, MPH, professor of psychiatry in the University of North Carolina (UNC) School of Medicine.

“Modern, computer-based methods can readily discover seemingly clear patterns from nonsensical data,” said Clark D. Jeffries, PhD, bioinformatics scientist at the UNC-based Renaissance Computing Institute (RENCI). “Added to that, scientific results from studies of complex disorders like schizophrenia can be confounded by many hidden dependencies. Thus, stringent testing is necessary to build a useful classifier. We did that.”

The study concludes that the multiplex blood assay, if independently replicated and if integrated with studies of other classes of biomarkers, has the potential to be of high value in the clinical setting.

Post a Comment

Your email address will not be published. Required fields are marked *