Researchers at the Rensselaer Polytechnic Institute (NY, USA) have developed an algorithm based upon the levels of a number of metabolites found in blood that could accurately predict whether a child is on the Autism spectrum, marking the first physiological test of its kind for autism.
Lead author Juergen Hahn, head of the Rensselaer Department of Biomedical Engineering, commented that “instead of looking at individual metabolites, we investigated patterns of several metabolites and found significant differences between metabolites of children with Autism spectrum disorder (ASD) and those that are neurotypical. These differences allow us to categorize whether an individual is on the Autism spectrum.”
Until now, the diagnosis of ASD is fully reliant on psychometric tests and behavioral observations, due to the lack of biological understanding of the cause. Therefore, the use of a physiological measure used in addition to the current methods of diagnosis could result in earlier and more accurate diagnosis.
In the study, the researchers analyzed the concentrations of metabolites present in blood that belong to two connected cellular pathways that are hypothesized to be associated with ASD: the folate-dependent one-carbon metabolism and transulfuration pathway.
The metabolite concentrations were measured from the blood samples of 83 patients with ASD and 76 age-matched controls. Using a Fisher Discriminant Analysis allowed for a multivariate nonlinear classification of participants with ASD being distinguished from controls.
Of the entire cohort, 97.6% of participants with ASD were correctly identified as being on the ASD spectrum and 96.1% of the control participants were correctly identified as not having ASD. To date, this accuracy in classification surpasses any other approach that has been used in this field.
Metabolites were also looked at on an individual basis, to determine if there were any links with ASD and a particular metabolite within the pathway; however, these results proved inconclusive. This highlights the need to use a more sophisticated technique, where many metabolites are studied and correlated with ASD.
It is hoped that the algorithm developed can be tested on other cohorts to determine whether these results can be replicated for conditions like Alzheimer’s or other neurodegenerative diseases.
Sources: Howsmon DP, Kruger U, Melnyk S, Jill James S, Hahn J. Classification and adaptive behavior prediction of children with autism spectrum disorder based upon multivariate data analysis of markers of oxidative stress and DNA methylation. PLoS Comput. Biol. 13(3) (2017); eng.rpi.edu/news/03162017-0000/blood-test-autism