An interpretable whole-blood transcriptomic model for Parkinson disease identification and molecular profiling
In the authors' words
Background Sporadic Parkinson disease (PD) lacks a molecular taxonomy that stratifies patients by dominant molecular drivers, limiting enrollment for disease-modifying trials, especially anti-inflammatory agents targeting NLRP3/NF-{kappa}B. Methods We developed ORIGIN-PD, an interpretable whole-blood transcriptomic classifier comprising 34 gene-level features and a composite low-expression score, with gene-level contributions organized into eight prespecified functional modules: midbrain dopaminergic (mDA) neuron identity/development, pathway-related, oxidative/mitochondrial, proteostasis, inflammation/ immunity, microglia, ncRNA, and other. Using LASSO regression on stringently quality-controlled whole-blood RNA-seq data from the Parkinson Progression Markers Initiative (PPMI) (Training Set: 204 PD and 92 controls; Locked Test Set: 51 PD and 23 controls), we assigned each patient a module-contribution profile. Module dominance was defined as >30% of total score. Results ORIGIN-PD discriminated PD from controls with high accuracy (Locked Test Set AUC 0.944, sensitivity 0.843, specificity 0.913). Among 255 sporadic PD patients, 49.4% exhibited a dominant module-contribution pattern, most frequently mDA neuron identity/development (22.0%), followed by pathway-related (11.8%), microglia/glial (7.5%), and inflammation/immunity (3.5%). Among the 50.6% with multi-module contributions, mDA neuron identity/development ranked in the top three in 62.8% of patients versus 35.7% for inflammation/immunity, indicating broader involvement of mDA neuron identity genes than inflammatory genes. The model showed high specificity against Alzheimer disease, ALS, and inflammatory controls. Conclusions ORIGIN-PD enables rational enrichment for NLRP3/NF-{kappa}B inhibitor trials by identifying patients with inflammatory involvement. In addition, the predominant contribution of the mDA neuron identity/development module, centred on WNT1, MSX1, NPM1, and LMX1B, identifies a larger patient subgroup with altered expression of mDA neuron identity module genes, suggesting that ORIGIN-PD may also be useful for patient enrichment in future trials of Wnt/{beta}-catenin pathway agonists. This dual-purpose framework provides a transparent, interpretable basis for pathogenic subtyping and precision enrichment in sporadic PD, with the aim of guiding disease-modifying therapy.
Appeared: Wednesday, October 21. medRxiv. Preprint, not yet peer-reviewed.