Parkinson’s disease is a neurodegenerative disorder that affects movement.
It currently has no cure and is typically diagnosed when physical symptoms such as hand tremors become noticeable.
However, scientists from UNSW Sydney and Boston University have recently made a significant stride towards the early detection of Parkinson’s disease, potentially years before these symptoms appear.
Using AI to Uncover Disease Biomarkers
In a study published in ACS Central Science, the researchers describe how they used machine learning – a form of artificial intelligence – to study biomarkers, or biological indicators of disease, in patients’ blood samples.
The team studied blood samples from healthy individuals, focusing particularly on 39 people who later developed Parkinson’s disease.
They ran a machine learning program over datasets that contained extensive information about metabolites, which are the chemical compounds our bodies produce when we break down food, drugs, or other substances.
CRANK-MS: A New Machine Learning Tool
The team developed a machine learning tool named CRANK-MS (Classification and Ranking Analysis using Neural network generates Knowledge from Mass Spectrometry) to analyze the metabolites.
Unlike traditional statistical approaches that focus on specific molecules, CRANK-MS takes into account the relationships between different metabolites, allowing a more comprehensive understanding of the body’s biochemical processes.
The researchers fed all the available data into CRANK-MS without initially reducing the number of chemical features, as is commonly done in other machine learning applications.
This approach enabled them to identify critical metabolites and determine their role in predicting the onset of Parkinson’s disease.
Implications for Parkinson’s Disease Diagnosis
Current methods for diagnosing Parkinson’s disease are limited to observing physical symptoms.
However, atypical symptoms such as sleep disorders and apathy can present decades before the motor symptoms of Parkinson’s disease appear.
This is where CRANK-MS could be a game-changer. The tool could be used early on, at the first sign of these atypical symptoms, to identify the risk of developing Parkinson’s disease in the future.
In the limited cohort examined for this study, the results were promising. CRANK-MS was able to analyze chemicals in blood samples to detect Parkinson’s disease with an impressive accuracy of up to 96 percent.
Dietary Habits and Parkinson’s Disease
The study also uncovered interesting findings regarding the diet of individuals who later developed Parkinson’s disease.
For instance, lower concentrations of triterpenoids, a known neuroprotectant found in foods such as apples, olives, and tomatoes, were found in the blood of those who later developed Parkinson’s disease.
Future studies could explore whether a diet rich in these foods could offer natural protection against the disease.
Another intriguing discovery was the presence of polyfluorinated alkyl substances (PFAS), which are industrial chemicals, in people who later developed Parkinson’s.
This finding suggests a possible link between exposure to certain chemicals and the onset of the disease.
A Tool Available for All
CRANK-MS is publicly available for any researchers interested in using machine learning for disease diagnosis using metabolomics data.
The tool’s design makes it fit for various applications and can generate results in less than 10 minutes on a regular laptop.
This is just one example of how AI can improve disease diagnosis and monitoring, potentially paving the way for early detection and better management of various diseases.
If you care about Parkinson’s disease, please read studies about Vitamin E that may help prevent Parkinson’s, and Vitamin D could benefit people with Parkinson’s.
For more information about brain health, please see recent studies about new way to treat Parkinson’s disease, and results showing flavonoid-rich foods could improve survival in Parkinson’s disease.
The study was published in ACS Central Science.
Copyright © 2023 Knowridge Science Report. All rights reserved.
