Qualitative analysis of broccoli (Brassica oleracea var. italica) glucosinolates: Investigating the use of mid-infrared spectroscopy combined with chemometrics

Glucosinolates are phytochemicals with important health and nutritional benefits. This study reports the use of high-performance liquid chromatography (HPLC) and mid-infrared (MIR) spectroscopy to characterise and differentiate between broccoli varieties and systems of production (organic vs. non-organic) depending on their glucosinolate content and infrared fingerprint. Broccoli samples (n = 53) from seven varieties were analysed using MIR spectroscopy and HPLC. Differences in the MIR spectra of the individual broccoli varieties were observed in the carbohydrate fingerprint region (950–1100 cm-1) and between 1340 and 1615 cm-1 assigned to specific glucosinolates. Principal component analysis (PCA) of the MIR fingerprint spectra enabled the differentiation between samples with relatively high (200–500 mg/100 g DW) and low (< 200 mg/100 g DW) glucobrassicin content. Linear discriminant analysis (LDA) and PCA-LDA were used to classify broccoli varieties according to the system of production (organic vs. non-organic) and variety (common vs. Tenderstem® broccoli). The classification rates indicated that > 70 % of the samples were correctly classified as organic and non-organic, while > 90 % of the samples were correctly classified as common broccoli and Tenderstem®. This study demonstrates that MIR spectroscopy could be used as a potential tool to classify and monitor broccoli samples according to their variety and system of production. •HPLC and MIR spectroscopy were used to characterise seven broccoli varieties.•PCA and LDA classification models were used on broccoli HPLC and MIR data.•MIR was able to differentiate between samples high or low in glucobrassicin content.•LDA of MIR correctly classified 96 % of samples as common broccoli or Tenderstem®.