Chicken, egg, or confounding effect? The need for deep phenotyping in psychiatric omics biomarker studies

Despite decades of research, there remain no replicable biomarkers for autism, not for most psychiatric disorders. More recently, high-throughput molecular “omics” technologies – that measure hundreds to millions of molecules at once – have become central in biomarker projects. Most omics biomarker studies focus on diagnostic associations, and essential questions of cause, consequence and confounding are often underexplored. Deeply phenotyped datasets are needed to address this, such as the Australian Biobank which has been my focus.

I will first discuss our microbiome study which provided evidence contrary to the much-hyped belief that microbiome plays a causal role in autism. Second, I show that lipidomics captures genetic and environmental interactions with autism and neurodevelopment, with particularly strong signal for sleep. Third, I will jump ship: from autism to cellular vulnerabilities that may contribute to Alzheimer’s disease risk and progression. I hope to demonstrate the importance of combining multi-omics studies with deep phenotyping and genetic data to help disentangle cause, consequence and confounding, and will discuss broader implications for autism biomarker research.

Join Zoom Meeting

Meeting ID: 944 5734 2362
Passcode: 051588