A quarterly publication of the Autism Research Institute
Summer, 2018 | Number 3, Volume 32
Researchers who reported last year on the results of a blood test for autism say that a new study confirms its efficacy.
The test, developed by Juergen Hahn and colleagues, involves analyzing 24 metabolites in blood samples in order to detect any significant differences between metabolites of children with ASD and neurotypical controls. These differences, the researchers say, allow them to predict whether an individual is on the autism spectrum.
In the initial study (see ARRI 31/2, 2017), the researchers analyzed data from 149 individuals, focusing on metabolites relevant to two cellular pathways linked to ASD: the methionine cycle and the transsulfuration pathway. About half of the children had ASD, while the other half were neurotypical. The researchers deliberately omitted data for one individual at a time, subjected the remaining data to advanced analysis techniques, and used the results to generate an algorithm to predict the data from the omitted individual. Repeating this process for all 149 children, the researchers correctly identified 96.1% of neurotypical children and 97.6% of children with ASD.
In the new study, the researchers used existing datasets from three studies involving a total of 154 children with autism. The datasets included only 22 of the metabolites used to create the original predictive algorithm. The researchers recreated their algorithm using data from the initial group of 149 children but limiting their analysis to these 22 metabolites. They then applied the algorithm to the new group of 154 children. This time, the algorithm predicted autism with 88% accuracy. Hahn says that the lower accuracy rate in the new study can most likely be attributed to the fact that two of the metabolites—both strong predictors in the initial study—were unavailable in this dataset.
“The most meaningful result.” Hahn says, “is the high degree of accuracy we are able to obtain using this approach on data collected years apart from the original dataset.”
“Multivariate techniques enable a biochemical classification of children with autism spectrum disorder versus typically-developing peers: A comparison and validation study,” Daniel P. Howsmon, Troy Vargason, Robert A. Rubin, Leanna Delhey, Marie Tippett, Shannon Rose, Sirish C. Bennuri, John C. Slattery, Stepan Melnyk, S. Jill James, Richard E. Frye, and Juergen Hahn, Bioengineering & Translational Medicine, June 2018 (open access). Address: Juergen Hahn, 110 Eighth St., Rensselaer Polytechnic Institute, CBIS #4213, Troy, NY 12180, [email protected].
—and—
“Success of blood test for autism affirmed,” news release, Rensselaer Polytechnic Institute, June 19, 2018.
—see also—
“Classification and adaptive behavior prediction of children with autism spectrum disorder based upon multivariate data analysis of markers of oxidative stress and DNA methylation,” Daniel P. Howsmon, Uwe Kruger, Stepan Melnyk, S. Jill James, and Juergen Hahn, PLOS Computational Biology, March 16, 2017 (open access). See address above.
