Supervised machine learning in the mass spectrometry laboratory: A tutorial.

Area of Science:

• Clinical Chemistry
• Computer Science
• Data Science

Background:

• Increasing demand for mass spectrometry (MS) testing necessitates automated data analysis.
• Machine learning (ML) methods are well-suited for structured clinical MS data, offering synergistic potential.
• A gap exists in technological literacy regarding ML software among clinical laboratory scientists.

Purpose of the Study:

• To provide a foundational overview of supervised ML principles.
• To outline the standard steps in an ML-based experiment.
• To discuss best practices for ML in the context of binary MS classification.

Main Methods:

• Demonstration of a supervised ML experiment using a published MS dataset.
• Explanation of fundamental supervised ML concepts.
• Tutorial on applying ML to a binary MS classification task.

Main Results:

• The study outlines the process of applying supervised ML to MS data.
• It highlights the potential for ML to optimize clinical laboratory workflows.
• The paper emphasizes the importance of good ML practices for reliable results.

Conclusions:

• Supervised ML presents a powerful tool for analyzing complex clinical MS data.
• Bridging the gap between clinical chemistry and computer science is crucial for effective ML implementation.
• This work serves as an educational resource for clinical scientists exploring ML applications in MS.