High-Resolution Separation Meets Data Fusion: A Quantitative Volatilomics Approach to Silage Evaluation

Silage, a fermented forage widely used in livestock feeding, undergoes microbial transformation under anaerobic conditions, with its quality and stability strongly influenced by the fermentation dynamics. In this context, volatilomics offers a powerful analytical approach to characterizing microbial metabolic activity and identifying markers related to product quality, spoilage, and stability.

In this study, the volatile profile of maize silage, both untreated and inoculated with a heterotactic bacterial strain, was analyzed after 100 days of conservation. The use of comprehensive two-dimensional gas chromatography (GC×GC) played a central role, as its superior separation capacity is crucial for resolving the highly complex chemical composition of volatilome matrices. In particular, differential flow-modulated GC×GC combined with parallel detection (mass spectrometry for structural elucidation and flame ionization detection for robust quantification) enabled detailed characterization and quantification of 98 volatiles across a wide concentration range. The adoption of multiple headspace SPME and predicted FID response factors allowed for accurate quantification without external calibration.

In addition to the discovery of several candidate markers, the complexity and dimensionality of the volatilomics data revealed the limitations of conventional, manually supervised workflows. To address this, a signal-level data fusion strategy was developed, integrating MS and FID outputs. The application of data fusion, driven by MS spectral similarity, minimizes feature mismatches, reducing false negatives when compared to those with FID alone and lowers false positives.

This integrated approach ensures confident quantification while preserving qualitative selectivity and facilitates the development of automated, high-throughput workflows. Applied to silage analysis, it enables efficient monitoring of the fermentation dynamics and stability over time.

Overall, the combination of GC×GC and data fusion enhances the analytical performance required for a marker-based quality assessment in complex biological matrices such as silage.