Analytica Chimica Acta 731 (2012) 24–31 Contents lists available at SciVerse ScienceDirect Analytica Chimica Acta jou rn al hom epa ge: www.elsevier.com/locate/aca Chemometric evaluation of different experimental conditions on wheat (Triticum aestivum L.) development using liquid chromatography mass spectrometry (LC–MS) proﬁles of benzoxazinone derivatives Mireia Farrés a , Marta Villagrasa b , Ethel Eljarrat a , Damià Barceló a,b , Romà Tauler a,∗ a Institute of Environmental Assessment and Water Research (IDAEA), Spanish Council for Scientiﬁc Research (CSIC), Jordi Girona 18-26, 08034 Barcelona, Spain b Catalan Institute for Water Research (ICRA), Parc Cientíﬁc i Tecnològic de la Universitat de Girona, Ediﬁci H20, Emili Grahit 100, 17003 Girona, Spain a r t i c l e i n f o Article history: Received 18 January 2012 Received in revised form 13 April 2012 Accepted 16 April 2012 Available online 21 April 2012 Keywords: Chemometrics Triticum aestivum L. Benzoxazinone derivatives LC–MS ANOVA–PCA ASCA a b s t r a c t Different chemometric techniques have been used to evaluate the effect of distinct experimental condi- tions and factors on Triticum aestivum L. plant development. The study was conducted using three wheat varieties, Astron, Ritmo and Stakado. These varieties were grown under organic and conventional cul- tivation systems. Samples were collected at ﬁve growth stages. Shoots and roots of each plant at these stages were analysed. Three replicates of each analysed sample were performed to improve representa- tiveness and to allow for the evaluation of natural variability and interaction effects. All samples were analysed using Liquid Chromatography Mass–Spectrometry (LC–MS), and the Total Ion Current (TIC) proﬁles of benzoxazinone derivatives obtained for each sample were investigated. Qualitative and quan- titative assessments of these TIC proﬁles and of their changes in the analysed samples were carried out using different chemometric techniques. Estimation of main effects, and of their possible interaction, was performed by means of Analysis of Variance combined to Principal Component Analysis (ANOVA–PCA) and of Analysis of Variance combined to Simultaneous Component Analysis (ASCA). © 2012 Elsevier B.V. All rights reserved. 1. Introduction In recent years, there has been an increasing interest to exploit- ing allelopathy as a biological strategy to minimize the perceived hazardous impacts from herbicides and insecticides in agriculture production. Allelopathy has been deﬁned as any direct or indi- rect effect (stimulatory or inhibitory) caused by a plant, including microorganisms, on another plant through production of secondary metabolites released into the environment [1,2]. The impact of allelopathy can be exploited for pest and weed control [3,4]. It has been proven that slight changes in the metabolism of plants can be explained by perturbations imposed on them (i.e. plants react to any change in their surroundings), and everything the plant does can be followed by looking at changes in the low molecular weight chemicals (metabolites). There is an enormous diversity of allelo- chemicals in nature [5]. Among them, benzoxazinones (hyrdoxamic acids, lactams, benzoxazolinones and methyl derivatives of hydrox- amic acids) are a group of secondary metabolites implicated on natural plant resistance. Benzoxazinones have been widely stud- ied during the last decade and several analytical methods using ∗ Corresponding author. Tel.: +34 93 400 61 40; fax: +34 93 204 59 04. E-mail address: Roma.Tauler@idaea.csic.es (R. Tauler). HPLC for these compounds have been developed [6,7]. Production of allelochemicals in living plants is affected by abiotic and biotic factors [8]. Understanding the complexity of the inﬂuence of different fac- tor levels in a plant growth using an experimental design strategy requires the use of multivariate data analysis methods. The inﬂu- ence of the factors of interest should be separated from each other to draw sensible conclusions from data analysis results. A widely used method for mutivariate data analysis is Principal Component Analysis (PCA). It gives a simpliﬁed lower-dimensional represen- tation of the variation that is presented in a dataset. The scores and loadings obtained by PCA can be visualised and interpreted in the context of the problem under study. However, this approach generally does not take the underlying experimental design into account. Thus, the different sources of variation are confounded in the PCA model, and this can seriously hamper the interpretation of the principal components [9]. A well recognised methodology for the analysis of data from designed experiments is Analysis of Variance (ANOVA), which focuses on the separation of the differ- ent sources of data variation. Therefore, combining ANOVA with PCA (ANOVA–PCA) will incorporate the pre-knowledge on the data structure into the model. This methodology has been already pro- posed [10] and it consists in using the experimental design to separate the variation of the experimental hypothesis from other 0003-2670/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.aca.2012.04.017