A simple blood test can now detect autism in children has been observed. An algorithm based on metabolites levels found in a blood sample can now predict accurately if a child is on the Autism spectrum of disorder (ASD), revealed a recent study. This is the foremost physiological test developed at Rensselaer Polytechnic Institute by researchers and it has opened the door to potential development of therapeutics in the future and also earlier diagnosis.
“Instead of considering the study of individual metabolites, the patterns were investigated on several metabolites and significant differences were found between metabolites of children with ASD and those that are neurotypical. Such differences helped in categorizing an individual is on the Autism spectrum,” said Juergen Hahn, lead author. “On measuring 24 metabolites from a blood sample, this algorithm tells whether or not an individual is on the spectrum of Autism, and even to some extent where on the spectrum they land can be ascertained.”
In the study, Hahn described Fisher Discriminant Analysis application, a big data analysis technique that comprises of data from a group of 149 people, about half on the Autism spectrum. Hahn subjected the dataset to advanced analysis techniques, and the results were used to generate a predictive algorithm. The algorithm makes a prediction about the omitted individual data. Hahn cross-validated the results, swapping different individuals out of the group and repeated the process in all 149 participants. His method identified correctly 96.1 percent neurotypical participants and 97.6 percent of the ASD cohort.
Hahn said, ‘ A lot of studies has been done taking one metabolite, one biomarker, one gene and there were differences found and most times they were not significant. Now it is done using big data techniques at a suite of metabolites and are found to be correlated with ASD, making the case stronger.