A comparison of semi-quantitative methods suitable for establishing volatile profiles

Victoria Ruiz-Hernández, María José Roca, Marcos Egea-Cortines, Julia Weiss

Full scent profiles emitted by living tissues can be screened by using total ion chromatograms generated in full scan mode and gas chromatography-mass spectrometry technique using Headspace Sorptive Extraction. This allows the identification of specific compounds and their absolute quantification or relative abundance. Quantifications ideally should be based on calibration curves using standards for each compound. However, the unpredictable composition of Volatile Organic Compounds (VOCs) and lack of standards make this approach difficult. Researchers studying scent profiles therefore concentrate on identifying specific scent footprints i.e. relative abundance rather than absolute quantities. We compared several semi-quantitative methods: external calibration curves generated in the sampling system and by liquid addition of standards to stir bars, total integrated peak area per fresh weight (FW), normalized peak area per FW, semi-quantification based on internal standard abundance, semi-quantification based on the nearest n-alkane and percentage of emission. Furthermore, we explored the usage of nearest components and single calibrators for semi-quantifications. Any of the semi-quantification methods based on a standard produced similar or even identical results compared to quantification by a true-standard for a compound, except for the method based on standard addition. Each method beholds advantages and disadvantages regarding level of accuracy, experimental variability, acceptance and retrieved quantities. Our data shows that, except for the method of standard addition to the biological sample, the rest of the semi-quantification methods studied give highly similar statistical results. Any of the methodologies presented here can therefore be considered as valid for scent profiling. Regarding relative proportions of VOCs, the generation of calibration curves for each compound analysed is not necessary.