The Aristotle Classifier: Using the Whole Glycomic Profile To Indicate a Disease State

Area of Science:

• Glycomics
• Bioinformatics
• Biotechnology

Background:

• Current glycomics strategies often focus on limited glycoforms or single biomarkers.
• This selective approach can overlook subtle yet significant changes in the overall glycomic profile.
• Ignoring the totality of glycan abundances and their relative proportions limits classification accuracy.

Purpose of the Study:

• To develop a novel classifier, the "Aristotle Classifier", that utilizes the entire glycomic profile for sample differentiation.
• To demonstrate a departure from traditional methods by incorporating thousands of glycomic features.
• To enhance the accuracy of classifying sample sources, including biological states like disease versus healthy.

Main Methods:

• Developed a classifier analyzing the totality of glycomic profiles, including all glycan abundances and their proportions.
• Applied the classifier to diverse glycomic data forms (derivatized monosaccharides, intact glycans, glycopeptides).
• Compared the Aristotle Classifier's performance against standard glycomics classification methods like Principal Component Analysis (PCA) and single-biomarker approaches.

Main Results:

• The Aristotle Classifier effectively differentiates samples from various sources by analyzing the complete glycomic landscape.
• This holistic approach captures subtle changes in glycan expression and proportions missed by traditional methods.
• The classifier demonstrated superior performance compared to PCA and single-biomarker workflows in multiple case studies.

Conclusions:

• The Aristotle Classifier offers a more robust and accurate method for sample classification in glycomics.
• Its ability to leverage the entire glycomic profile makes it highly effective for biomarker discovery and source attribution.
• This approach represents a significant advancement in analyzing complex glycomic data for biological and biotechnological applications.