Machine learning-guided spectral fingerprinting reveals Parkinson's disease-associated biochemical changes in Drosophila melanogaster

Fuente: PubMed "medicinal and aromatic plants"
Spectrochim Acta A Mol Biomol Spectrosc. 2026 Aug 4;364(Pt 1):128563. doi: 10.1016/j.saa.2026.128563. Online ahead of print.ABSTRACTParkinson's disease (PD) is a neurodegenerative disorder characterized by degeneration of dopaminergic neurons, oxidative stress, mitochondrial dysfunction, and metabolic imbalance. These pathological processes alter cellular biochemical composition and may generate detectable spectral signatures. In the present study, Fourier transform infrared (FT-IR) spectroscopy was employed to investigate biochemical alterations associated with neurotoxicant-induced PD phenotypes in Drosophila melanogaster. PD-like phenotypes were induced using paraquat, rotenone, and 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) and were validated through locomotor impairment and dopaminergic neuronal degeneration. Spectra (400-4000 cm-1) revealed characteristic alterations within lipid (2800-3000 cm-1), protein amide (∼1650 and ∼ 1540 cm-1), and fingerprint regions indicating modifications in macromolecular composition following neurotoxic exposure. Principal component analysis (PCA) and hierarchical clustering revealed intrinsic spectral differences between control and treated groups, while partial least squares discriminant analysis (PLS-DA) supported group discrimination, with permutation testing indicating that the observed separation was unlikely to arise by chance. Variable importance in projection (VIP) analysis identified prominent contributions from fingerprint and lipid-associated spectral regions, particularly within the 1012-1040 cm-1 domain. Classification models (KNN, RF) further supported robust separation between control and PD groups using independent test datasets. The observed spectral alterations were consistent with independent biological evidence of oxidative stress, mitochondrial dysfunction, lipid accumulation, and neurodegeneration. Together, these findings demonstrate that ATR-FTIR spectroscopy provides a rapid, label-free, and biologically informative approach for characterizing PD biochemical alterations and highlight its potential application in neurodegenerative disease research and neurotoxicity assessment.PMID:42585923 | DOI:10.1016/j.saa.2026.128563