Scientists in Sydney, New South Wales, working with collaborators at Boston University, have developed a tool that shows the potential for early detection of Parkinson’s disease, years before the first symptoms appear.
In a study published today in the journal ACS Central Scienceresearchers described how they used neural networks to analyze biomarkers in patient fluids.
Researchers at the New South Wales School of Chemistry tested blood samples taken from healthy individuals collected by the European Prospective Investigation on Cancer and Nutrition (EPIC) in Spain. The research team focused on 39 patients who developed Parkinson’s disease by 15 years and included extensive information on metabolites (compounds the body produces when breaking down foods, drugs, or chemicals). I ran a machine learning program on the dataset.
Want more breaking news?
apply technology network‘ Daily Newsletter. Get the latest science news delivered directly to your inbox every day.
subscribe for free
Comparing these metabolites with metabolites in 39 control patients in the same study who did not develop Parkinson’s disease may prevent or be an early warning sign of Parkinson’s disease. We were able to identify a unique combination of metabolites.
As UNSW researcher Diana Zhang explains, she and Associate Professor W. Alexander Donald developed a machine learning tool called CRANK-MS. It stands for Classification and Ranking Analysis Using Neural Networks to Generate Knowledge from Mass Spectrometry.
“The most common way to analyze metabolomics data is through statistical approaches,” says Zhang.
“Thus, to figure out which metabolites are more important for diseased and control groups, researchers usually look for correlations on specific molecules.
“But here we take into account that metabolites can be related to other metabolites. This is where machine learning comes in. We can use hundreds or thousands of metabolites to figure out what I used my computational power to understand what was going on.”
A/Prof. Donald said the researchers used a list of unedited data in addition to looking at combinations of metabolites.
“Typically, researchers using machine learning to look at correlations between metabolites and diseases first reduce the number of chemical features before feeding them into the algorithm,” he says.
“But here we don’t do any data reduction to begin with, we just feed all the information into CRANK-MS. And from there we get the model predictions, and we all know which metabolites are driving the predictions the most.” step, which means we can now detect metabolites that may have been missed by traditional approaches.”
How this is important for Parkinson’s disease
Parkinson’s disease is currently diagnosed by observing physical symptoms such as hand tremors at rest. There are no blood tests or laboratory tests to diagnose non-genetic cases. However, people with Parkinson’s disease may develop atypical symptoms, such as sleep disturbances and apathy, decades before motor symptoms appear. Therefore, CRANK-MS may be used at the first signs of these atypical symptoms to rule out or eliminate the risk of developing Parkinson’s disease in the future.
However, Professor Donald stresses that validation studies using much larger cohorts and conducted in multiple regions of the globe are needed to ensure the tool’s use. However, in the limited cohort investigated in this study, CRANK-MS was able to analyze chemicals in blood and detect Parkinson’s disease with up to 96% accuracy, showing encouraging results. rice field.
“This study is interesting on many levels,” he says.
“First, the accuracy of predicting Parkinson’s disease prior to clinical diagnosis is very high. Third, some of the chemical markers that drive the most accurate predictions have been previously implicated in Parkinson’s disease by others in cell-based assays. but not in humans.”
food for thought
When the researchers looked at metabolites in people who developed Parkinson’s disease, they made some interesting discoveries.
For example, people who later developed Parkinson’s disease had lower blood triterpenoid levels than those who did not. Triterpenoids are known neuroprotective substances that regulate oxidative stress and are commonly found in foods such as apples, olives and tomatoes. Future research may investigate whether eating these foods can naturally prevent the development of Parkinson’s disease.
The presence of polyfluoroalkyl substances (PFAS) in people who develop Parkinson’s disease also deserves further study and may be associated with exposure to industrial chemicals.
“There is evidence to suggest it is PFAS, but we need more characterization data to be 100% sure,” says A/Prof. Donald.
Free for everyone
CRANK-MS is a publicly available tool for researchers wishing to use machine learning for disease diagnosis using metabolomics data.
“We built the model in a fit-for-purpose manner,” says Zhang.
“The application of CRANK-MS to detect Parkinson’s disease is just one example of how AI can improve the way we diagnose and monitor disease. It can be easily applied to identify new biomarkers of interest.
“The tool is easy to use, and on average we get results in less than 10 minutes on a traditional laptop.”
reference: Zhang JD, Xue C, Kolachalama VB, Donald WA. Interpretable machine learning on metabolomics data reveals biomarkers for Parkinson’s disease. ACS Central Science. 2023.doi: 10.1021/accentsci.2c01468
This article is reprinted from: Note: Materials may have been redacted for length and content. Please contact the citation source for details.
