Multiset data analysis in mixture designed fingerprints through multivariate curve resolution: exploring spectroscopy data fusion approaches in metabolomics

• Grant number: 22/12465-7
• Support Opportunities: Scholarships abroad - Research Internship - Post-doctor
• Start date: May 01, 2023
• End date: April 30, 2024
• Field of knowledge: Physical Sciences and Mathematics - Chemistry - Analytical Chemistry
• Principal Investigator: Roy Edward Bruns
• Grantee: Gustavo Galo Marcheafave
• Supervisor: Roma Tauler
• Host Institution: Instituto de Química (IQ). Universidade Estadual de Campinas (UNICAMP). Campinas , SP, Brazil
• Institution abroad: Instituto de Diagnóstico Ambiental y Estudios del Agua (IDAEA), Spain
• Associated to the scholarship: 20/11463-5 - Extractor solvent-detector optimization of plant responses to environmental impacts: an integrated factorial design, mixture design and ANOVA: simultaneous component analysis strategy, BP.PD

Abstract

Metabolomic studies using multiset data analysis approaches have been scarcely used in Brazil until now. No instrumental method of chemical analysis or of sample preparation is capable to evaluate and extract the entire metabolome of a biological system. Some chemometric methods have been proposed to simultaneously analyze multiple metabolomic data sets, such as multivariate curve resolution-alternating least squares (MCR-ALS). MCR-ALS decomposes a data matrix, D, by a bilinear model into the product of two factor matrices that describes the set of concentration, C, and spectral profiles, ST, of the analyzed systems. Then, the resolved concentration data can be arranged (SuperC matrix) and explored by multivariate chemometric methods to discriminate, classify, and determine the effects and interactions of the studied factors in designed experiments. Moreover, the information contained in the spectral data matrix can be used for the identification of the chemical (metabolite) components present in the analyzed systems. In this project, different extended type MCR models will be evaluated using row- and column-wise augmented experimentally designed data matrices individually and simultaneously. For this, mixture design-fingerprints of yerba-mate leaves acquired by four instrumental techniques from a 3-component statistical mixture design will be analyzed in detail. This procedure intends to evaluate the quantitative behavior of chemical species in relation to the extractor system and the simultaneous resolution of diverse instrumental spectra of the constituents of the analyzed samples with the goal of making the metabolic analysis more robust and globally inclusive. This knowledge will contribute significantly to the improvement of the analysis skills of national research groups that are applying chemometric tools in metabolic analysis at this time. (AU)