Untargeted Chemical Profiling of Two-Dimensional Gas Chromatography Coupled with High-Resolution Mass Spectrometry Data for Botrytized Wines via Topological Data Analysis

Advanced chemical profiling of complex samples, such
as botrytized
wines, requires advanced analytical techniques capable of capturing
subtle compositional variations. In this study, we introduce a statistically
robust framework that leverages a topological data analysis (TDA)
tool, Ball Mapper, in the context of comprehensive two-dimensional
gas chromatography (GC × GC) with high-resolution time-of-flight
mass spectrometry (HR-TOF-MS) to obtain untargeted identification
of sample-specific chemical markers. A key design element of the proposed
approach is its ability to numerically process the immense data volume
generated per sample, whose statistical and chemical significance
is often difficult to interpret using conventional methods. Each of
the 34 wine samples yielded over 470,000 mass spectral functions,
which were discretized, normalized, and clustered to obtain representative
and relatively unique discrete mass spectral vectors in high-dimensional
space. With only two interpretative parameters, the proposed framework
uncovered 2,792 extracted mass spectral distributions, from which
1191
discriminative features were identified, including 334 compounds assigned
to known volatile organic compound classes. The resulting chemical
signatures reflected regional differences in fermentation style, grape
variety, botrytization conditions, and microbial activity. Moreover,
statistically robust framework of using Ball Mapper revealed consistent
grouping patterns both within and between wines. These findings demonstrate
that the proposed framework can support chemical characterization
complex natural matrices and serve as a general strategy for analyzing
any domains where GC × GC with HR-TOF-MS data are collected.