Integration of data from GC-MS and UPLC-QTOF-MS to better understand wine ageing

Entre em https://eu.bbcollab.com/guest/d28307c5ab554234afe3dfbb16736d94 no dia 29 de janeiro de 2021 às 14h15 e assista à palestra de António César Ferreira.

A "entrada" é livre e gratuita, sem necessidade de pré-inscrição.

Given the global trends in the food and flavor industry, i.e. naturality, organic food, authenticity, transparency as well as ’clean label’, it has become essential to revisit the way flavor is imparted in foods and beverages, as well as the way it is measured with fast reproducible methods, close to the overall consumer experience. In the last decade, there has been an increase in analytical methods allowing the acquisition of volatile and non-volatile fractions of different matrices like GC and LC-MS. The application purpose of this technique includes a variety of projects from origin and history assessment of food and raw products, to food quality monitoring, to flavor release and/or generation profiles as function of product composition, to sensory related studies. However, the limitation of these high-throughput techniques remains with the considerable amount of data generated in a single analysis, in a complex data structure, and the required prior number of steps to a robust analysis of the acquired information. Hence, the need to develop alternative data treatment concepts, capable to extract relevant information, supporting the usage of fast real time measurements, close to the overall consumer experience. The present talk will elaborate on a few of the these current issues and on how the food industry is transforming challenges into opportunities, with the development of a pipeline, consisting on: data importation, peak apex’s extraction to reduce the number of variables and integrating several statistic tools for data comparison and visualization. In order to have a holistic view of the chemical system a pipeline was developed based on UPLS-MS-QTOF and GC-MS data acquisition followed by data fusion. The process is hyphenated with an in-house peak picking interface, coupled with multi- and -univariate statistics to get the most relevant compounds.