ADVANCING FECAL VOLATILOME PROFILING BY TWO-DIMENSIONAL GAS CHROMATOGRAPHY-TIME OF FLIGHT MASS SPECTROMETRY (GC×GC-TOF MS) AND IMAGE PATTERN RECOGNITION

Comprehensive two-dimensional gas chromatography-time of flight mass spectrometry (GC×GC-TOF MS) represents a significant advancement in analytical techniques, offering exceptional resolution, sensitivity and the ability to characterize complex samples with high chemical dimensionality. Unlike one-dimensional gas chromatography methods, GC×GC-TOF MS provides a more precise and detailed fingerprint of volatile organic compounds (VOCs) in a closed range of volatilities/polarities where multiple co-elution might occur. Moreover, thanks to the band-compression in space provided by thermal modulation, a very large dynamic range of concentrations can be covered. This study investigates the potential of GC×GC-TOF MS for the comprehensive analysis of fecal volatile metabolites, focusing on its application in profiling the fecal volatilome, an area that remains underexplored in metabolomics. Fifty individuals with suspected NCGS/WS were subjected to the double blind-placebo-controlled crossover gluten challenge test. Twenty-seven participants were gluten responsive (NCGS) and randomized into two arms: Experimental arm: 6 weeks with daily intake of commercial probiotics (L. plantarum, L. paracasei and L. salivarius) or Control arm: 6 weeks with daily intake of placebo. All participants were undergoing a gluten-free diet for 4 weeks (sampling point T1), after which gluten was reintroduced for 2 weeks (sampling point T2). By applying combined untargeted and targeted fingerprinting based on image pattern recognition on GC×GC-TOF MS data, diagnostic signatures of the diet intervention were highlighted. Over the 830 UT features (i.e., detectable volatilome), consistently aligned across samples 2D patterns, about 200 were putatively identified and correlated to the study variables by both unsupervised and supervised chemometrics. Results confirmed meaningful changes in the fecal volatilome after 4 weeks (T1) of pro-biotic treatment (vs. placebo) accompanied by a gluten-free diet consolidated after further 2 weeks of treatment (T2) and normal diet. Partial-Least Squares Discriminant Analysis (PLS-DA) provided good classification models with 89% accuracy at T1 4 weeks and 90% at T2 6 weeks (considering both UT features % Response and Absolute Response). Targeted features with higher information potential resulted butanoic and propanoic acids esters, primary alcohols, aldehydes and several terpenoids. By Computer Vision (CV) image patterns were highlighted and confounding variables minimized between samples classes. Although based on a limited samples set, the comprehensive mapping of fecal volatilome by GC×GC, provides essential information on metabolic changes induced by diet intervention and pro-biotic supplementation supporting personalized nutrition. Moreover, by extending the knowledge on fecal volatilome composition, the complexities of biological systems can be better elucidated by connecting information with other omics (e.g., metagenomics).