Development of Diagnostic Tests for Autism-Spectrum Disorders: Data Science Meets Metabolomics

Lecture / Panel
For NYU Community

Autism Ribbon


Juergen Hahn, PhD
Professor and Chair, Dept. of Biomedical Engineering
Rensselaer Polytechnic Institute (RPI)


Autism Spectrum Disorders (ASD) are a group of neurological disorders that present with limited social communication/interaction and restricted, repetitive behaviors/interests. The current estimate is that approximately 1.9% of children in the US are diagnosed with ASD. While this is a high prevalence and the economic burden by ASD is significant, there is still considerable debate regarding the underlying pathophysiology of ASD. Because of this lack of biological knowledge, autism diagnoses are restricted to observational behavioral and psychometric tools. This work takes a step towards the goal of incorporating biochemical data into ASD diagnosis by analyzing measurements of metabolite concentrations of the folate-dependent one-carbon metabolism and transulfuration pathways. Unlike traditional approaches that are based upon comparing differences in individual metabolite concentrations between children with and without an ASD diagnosis, we made use of multivariate classification via Fisher Discriminant Analysis and used Kernel Partial Least Squares regression to predict adaptive behavior. The resulting levels of accuracy for classification as well as adaptive behavior prediction exceed any other approach in this field.

Dr. Hahn received his Diploma degree in engineering from the Technical University Aachen in Germany in 1997 and his MS and PhD degrees in chemical engineering from UT Austin. He joined the Department of Chemical Engineering at Texas A&M University, College Station, in 2003 and moved to the RPI in 2012. Dr. Hahn is a recipient of a Fulbright scholarship, was 4-times named outstanding reviewer by the journal Automatica, and has been elected as an AIMBE Fellow in 2013 and an AIChE Fellow in 2020. He served on the IEEE CSS Board of Governors in 2016 and has been a CACHE Trustee since 2014. Currently he is the deputy editor-in-chief for the Journal of Process Control, the editor for the journal Optimal Control: Applications and Methods, and associate editor for Control Engineering Practice and the Journal of Advanced Manufacturing and Processing.

Example of correlation-based network analysis with plasma metabolites
Example of correlation-based network analysis with plasma metabolites associated with clinical assessment scores (highlighted in green). Blue and red lines between metabolites indicate positive and negative correlations, respectively (Spearman correlation coefficient |r| > 0.5).