Multi-omic platform for identifying causal and druggable alzheimer's disease targets

Overview

This research program delivers a comprehensive multi-tissue, multi-omic platform designed to unravel the molecular architecture of Alzheimer's disease (AD) and accelerate the discovery of causal biomarkers and druggable targets. By integrating genomics, transcriptomics, epigenomics, proteomics, metabolomics, and lipidomics data from brain, cerebrospinal fluid (CSF), and plasma samples drawn from large, well-characterized cohorts, the platform enables deep molecular profiling that has historically transformed precision medicine in oncology but remains underutilized in neurodegenerative disease research.

The approach applies advanced analytical methods, including Mendelian Randomization, colocalization, and machine learning, to identify causal genes, proteins, and analytes implicated in AD risk and progression. Results to date include the identification of one CSF protein, thirteen plasma proteins, and six brain proteins in likely causal AD pathways, along with several druggable targets matched to FDA-approved compounds suitable for repositioning. Prediction models built from CSF and plasma data outperform the current gold-standard CSF amyloid-beta/phosphorylated tau biomarker, and novel models have been created for TREM2 variant carriers and individuals with autosomal dominant AD mutations.

Technical specifications

Key capabilities:

• Multi-omic profiling across brain, CSF, and plasma tissues using next-generation sequencing, transcriptomics, methylation arrays, proteomics (Somalogic 1.3k platform), metabolomics, and lipidomics
• Protein quantitative trait locus (pQTL) mapping to link genetic variation to protein abundance
• Mendelian Randomization and colocalization analyses to distinguish causal from correlated molecular signals
• Machine learning prediction models for sporadic AD, TREM2 carriers, and autosomal dominant AD mutation carriers
• Drug repositioning pipeline identifying FDA-approved compounds that target causal AD proteins
• Tissue-specific molecular signature profiling for sporadic and genetically defined AD subtypes

Validation evidence:

• Proteomic data generated from 971 CSF samples, 636 plasma samples, and 458 brain samples
• Hundreds of novel pQTL signals identified across tissues
• Prediction models in CSF and plasma demonstrating superior performance over gold-standard CSF amyloid-beta/phosphorylated tau ratios
• First-ever prediction models developed for TREM2 and autosomal dominant AD mutation carriers

Technology readiness level

The platform has generated and analyzed initial proteomic datasets across three tissue types, with peer-reviewed results confirming causal protein identification and superior biomarker performance. Current work is scaling the approach to a multi-omic QTL atlas spanning all planned tissue and molecular layers. Future validation aims include generating complete multi-omic datasets in large cohorts, building tissue-specific prediction models, and systematically cataloging druggable targets for repositioning. The technology is positioned for translational partnerships with pharmaceutical and biotechnology organizations seeking validated causal targets and biomarker panels for Alzheimer's disease drug development.