Multivariate curve resolution for high-throughput carotenoid profiling

Carotenoids are usually not considered in metabolomic approaches due to their instability, hydrophobicity and the difficulties in analyzing them by MS. A targeted high-performance liquid chromatography with diode array detection (HPLC-DAD) method (Fraser, 2000) has been used in several metabolomic approaches but data analysis of relatively large sample numbers traditionally involves the manual integration of hundreds of peaks at different wavelengths. Some degree of automation is allowed by data analysis software provided with chromatographic instruments, but in practice several problems are encountered and the process is cumbersome and error prone. Here a multivariate curve resolution (MCR) approach is proposed as a method to analyse large numbers of HPLC-DAD data sets. A case study in which the stability of carotenoids in grape extracts processed in the presence or absence of triethylamine (TEA) is studied is used to prove the capacities of the method