Alignment and clustering strategies for GC×GC-MS features using a cylindrical mapping.

Area of Science:

• Analytical Chemistry
• Chemometrics
• Biomarker Discovery

Background:

• Comprehensive two-dimensional gas chromatography coupled to mass spectrometry (GC×GC-MS) is vital for analyzing complex samples.
• Extensive data preprocessing, including alignment, is crucial for large-scale studies like biomarker discovery.
• Existing alignment methods face challenges with phenomena like wrap-around.

Purpose of the Study:

• To develop novel alignment and clustering algorithms for GC×GC-MS data.
• To address the wrap-around phenomenon in GC×GC-MS data analysis.
• To enhance the robustness and repeatability of GC×GC-MS data analysis.

Main Methods:

• Developed new alignment and clustering algorithms for GC×GC-MS data.
• Treated the 2D data display as a 3D cylinder surface to handle wrap-around.
• Introduced a new similarity metric based on cylindrical distance and mass spectral correlation.
• Incorporated protection against greedy warping in the algorithms.

Main Results:

• The new algorithms effectively align complex GC×GC-MS data, addressing the wrap-around issue.
• The alignment process demonstrated robustness against greedy warping.
• Chemically similar features were successfully clustered together.
• Applied methods to analyze replicates of human urine samples.

Conclusions:

• The developed algorithms provide a robust and consistent approach for GC×GC-MS data alignment and clustering.
• The cylindrical transformation and new similarity metric improve data analysis accuracy.
• These advancements are beneficial for applications such as biomarker discovery, enhancing analytical quality.