A multi-dimensional study on taste deficits, α-synuclein genetic profiling, and supervised learning for diagnosis and disease severity stratification in Parkinson's disease (PD)

Parkinson's disease (PD) is a complex neurological condition. Alfa-synuclein, and specifically the salivary oligomeric form, have been described as biomarkers of PD, and their genetic mutations have been associated with PD susceptibility. Taste impairments are less studied and have produced inconsistent results, and their diagnostic and prognostic significance remains underexplored. We evaluated taste perceptions for five taste stimuli and their association with four α-synuclein gene (SNCA) polymorphisms in PD patients and healthy controls (HCs). Demographic and clinical features of the PD patients were also determined. A Supervised Learning (SL) model classified PD versus HC and moderate versus high severity PD (MSPD versus HSPD), determining the impact of each feature. We found that taste impairments in PD are modality-specific, with saltiness and astringency most affected, and are modulated by the SNCA gene variations. NaCl under threshold, age, and incorrect identification of tannic acid were the most influential features for PD prediction. Specific genotypes of SNCA SNPs (e.g., rs356219 GG, rs181489 TT, rs2583988 CC, rs356186 GG) were enriched in PD patients with impaired astringency and saltiness. Tremor-dominant PD and shorter PD were associated with MSPD, while longer disease duration and akinetic-rigid type predicted HSPD. The rs2583988 CC genotype was associated with moderate PD, while the rs2583988 TT genotype and lower taste acuity predicted high severity. Our findings showed that, when combined with SL and genetic profiling, these sensory markers offer a powerful, non-invasive approach for early diagnosis and disease stratification, supporting the integration of sensory and genetic screening into clinical practice.