ADSP Phenotype Harmonization Consortium
Timothy J. Hohman- Psychiatry and Mental health
- Cellular and Molecular Neuroscience
- Geriatrics and Gerontology
- Neurology (clinical)
- Developmental Neuroscience
- Health Policy
- Epidemiology
Abstract
Background
The Alzheimer’s Disease Sequencing Project Phenotype Harmonization Consortium (ADSP‐PHC) was assembled in 2021 to provide large‐scale harmonization of 40+ deeply phenotyped cohorts included in the ADSP. Harmonization spans from markers of amyloid, tau, neurodegeneration, concomitant pathways of injury, and cognitive decline. This presentation will provide an overview of the methods and results of our first data release (October 2022) and provide a preview for upcoming data release in the coming years.
Method
A team of domain experts was established to provide statistical harmonization. The outcome of these specialized harmonization procedures includes longitudinal composite measures of executive function, language, and memory performance leveraging latent variable modeling, longitudinal biomarker measurement of Aβ42, total tau, and phosphorylated tau leveraging a distribution based harmonization, longitudinal amyloid positron emission tomography (PET) harmonization accounting for different ligands and scanner protocols across sites, longitudinal magnetic resonance imaging (MRI) harmonization including diffusion imaging leveraging image‐based harmonization and statistical harmonization with COMBAT, and both vascular risk factor and neuropathology harmonization leveraging a detailed variable mapping approach. All data, scripts, and documentation are made freely available to qualified investigators on dss.niagads.org.
Result
The first data release (October of 2022) and is described in detail here:
Conclusion
The deeply phenotyped cohort studies within the ADSP provide an unprecedented opportunity to explore the genetic architecture of amyloid, tau, neurodegeneration, and cognitive decline. The ADSP‐PHC will deliver an unparalleled resource of harmonized data across domains to facilitate novel genomic discovery and provide AI‐ready resources to better understand drivers of disease heterogeneity.