Neurodegenerative diseases (NDs) are multifactorial and currently lack scalable, non-invasive tools for early and preclinical diagnosis, relying instead on costly or invasive methods such as CSF sampling and PET imaging. To address this gap, we developed NeuroDeGene®, a DNA-targeted sequencing panel (created with Celemics) designed to detect SNPs and INDELs across 19 neurodegenerative disorders. The panel comprises 15,533 variants curated from ClinVar, GWAS, and PGS catalogs, mapped to 5,219 genes, with genome-wide coverage and over 70% non-coding variants.

Compared to whole-genome sequencing, NeuroDeGene® offers higher coverage at targeted capture regions, greater median coverage for selected variants, and a lower error rate (~1%), at reduced cost. Validation in the ADDIA cohort (NCT03030586) showed that ML-based classification of Alzheimer's disease (AD) versus healthy controls captured ~75% of the genetic signal, consistent with published AD heritability estimates. While the APOE locus remains the strongest single genetic risk factor, machine learning models incorporating the broader panel improved differentiation of AD from other dementias, including dementia with Lewy bodies (DLB), beyond what APOE alone provides.

These results support NeuroDeGene®'s potential for early genetic risk stratification and differential diagnosis across neurodegenerative diseases. We are actively seeking access to well-characterized neurodegenerative disease cohorts to further expand and validate the panel for multi-disease differential diagnosis.