Combining Machine Learning and Metabolomics to Identify Weight Gain Biomarkers

Area of Science:

• Metabolomics
• Biomarker Discovery
• Obesity Research

Background:

• Weight gain is a metabolic disorder often leading to obesity and related conditions like diabetes.
• Obesity is associated with chronic, subclinical systemic inflammation, a significant contributor to comorbidities.
• Identifying individuals at high risk for severe weight-associated morbidity is crucial for public health.

Purpose of the Study:

• To identify and characterize biomarkers associated with weight gain in a Brazilian population.
• To analyze plasma samples for metabolic signatures related to weight gain and inflammation.
• To leverage machine learning for biomarker discovery in obesity.

Main Methods:

• Plasma samples from 180 Brazilian individuals (eutrophic and case groups) were analyzed.
• Mass spectrometry was employed for comprehensive sample analysis.
• The Random Forest machine learning algorithm was used to identify discriminant features.

Main Results:

• Five biomarkers related to weight gain pathogenesis and inflammation were identified.
• Upregulated arachidonic acid metabolites indicated inflammation.
• Dysfunctions in nitric oxide (NO) cycle and increased superoxide production were observed.
• A novel marker potentially linked to diabetes onset in overweight/obese individuals was found.

Conclusions:

• Mass spectrometry combined with machine learning effectively identified weight gain biomarkers.
• The identified biomarkers offer potential for improved risk assessment and characterization of obesity phenotypes.
• These findings may pave the way for novel therapeutic and prognostic targets for weight-associated diseases.