Thèses sur le sujet « Chemometric approaches »

The present Ph.D. thesis is devoted to study, develop and apply approaches commonly used in chemometrics to the emerging field of systems biology. Existing procedures and new methods are applied to solve research and industrial questions in different multidisciplinary teams. The methodologies developed in this document will enrich the plethora of procedures employed within omic sciences to understand biological organisms and will improve processes in biotechnological industries integrating biological knowledge at different levels and exploiting the software packages derived from the thesis. This dissertation is structured in four parts. The first block describes the framework in which the contributions presented here are based. The objectives of the two research projects related to this thesis are highlighted and the specific topics addressed in this document via conference presentations and research articles are introduced. A comprehensive description of omic sciences and their relationships within the systems biology paradigm is given in this part, jointly with a review of the most applied multivariate methods in chemometrics, on which the novel approaches proposed here are founded. The second part addresses many problems of data understanding within metabolomics, fluxomics, proteomics and genomics. Different alternatives are proposed in this block to understand flux data in steady state conditions. Some are based on applications of multivariate methods previously applied in other chemometrics areas. Others are novel approaches based on a bilinear decomposition using elemental metabolic pathways, from which a GNU licensed toolbox is made freely available for the scientific community. As well, a framework for metabolic data understanding is proposed for non-steady state data, using the same bilinear decomposition proposed for steady state data, but modelling the dynamics of the experiments using novel two and three-way data analysis procedures. Also, the relationships between different omic levels are assessed in this part integrating different sources of information of plant viruses in data fusion models. Finally, an example of interaction between organisms, oranges and fungi, is studied via multivariate image analysis techniques, with future application in food industries. The third block of this thesis is a thoroughly study of different missing data problems related to chemometrics, systems biology and industrial bioprocesses. In the theoretical chapters of this part, new algorithms to obtain multivariate exploratory and regression models in the presence of missing data are proposed, which serve also as preprocessing steps of any other methodology used by practitioners. Regarding applications, this block explores the reconstruction of networks in omic sciences when missing and faulty measurements appear in databases, and how calibration models between near infrared instruments can be transferred, avoiding costs and time-consuming full recalibrations in bioindustries and research laboratories. Finally, another software package, including a graphical user interface, is made freely available for missing data imputation purposes. The last part discusses the relevance of this dissertation for research and biotechnology, including proposals deserving future research.
Esta tesis doctoral se centra en el estudio, desarrollo y aplicación de técnicas quimiométricas en el emergente campo de la biología de sistemas. Procedimientos comúnmente utilizados y métodos nuevos se aplican para resolver preguntas de investigación en distintos equipos multidisciplinares, tanto del ámbito académico como del industrial. Las metodologías desarrolladas en este documento enriquecen la plétora de técnicas utilizadas en las ciencias ómicas para entender el funcionamiento de organismos biológicos y mejoran los procesos en la industria biotecnológica, integrando conocimiento biológico a diferentes niveles y explotando los paquetes de software derivados de esta tesis. Esta disertación se estructura en cuatro partes. El primer bloque describe el marco en el cual se articulan las contribuciones aquí presentadas. En él se esbozan los objetivos de los dos proyectos de investigación relacionados con esta tesis. Asimismo, se introducen los temas específicos desarrollados en este documento mediante presentaciones en conferencias y artículos de investigación. En esta parte figura una descripción exhaustiva de las ciencias ómicas y sus interrelaciones en el paradigma de la biología de sistemas, junto con una revisión de los métodos multivariantes más aplicados en quimiometría, que suponen las pilares sobre los que se asientan los nuevos procedimientos aquí propuestos. La segunda parte se centra en resolver problemas dentro de metabolómica, fluxómica, proteómica y genómica a partir del análisis de datos. Para ello se proponen varias alternativas para comprender a grandes rasgos los datos de flujos metabólicos en estado estacionario. Algunas de ellas están basadas en la aplicación de métodos multivariantes propuestos con anterioridad, mientras que otras son técnicas nuevas basadas en descomposiciones bilineales utilizando rutas metabólicas elementales. A partir de éstas se ha desarrollado software de libre acceso para la comunidad científica. A su vez, en esta tesis se propone un marco para analizar datos metabólicos en estado no estacionario. Para ello se adapta el enfoque tradicional para sistemas en estado estacionario, modelando las dinámicas de los experimentos empleando análisis de datos de dos y tres vías. En esta parte de la tesis también se establecen relaciones entre los distintos niveles ómicos, integrando diferentes fuentes de información en modelos de fusión de datos. Finalmente, se estudia la interacción entre organismos, como naranjas y hongos, mediante el análisis multivariante de imágenes, con futuras aplicaciones a la industria alimentaria. El tercer bloque de esta tesis representa un estudio a fondo de diferentes problemas relacionados con datos faltantes en quimiometría, biología de sistemas y en la industria de bioprocesos. En los capítulos más teóricos de esta parte, se proponen nuevos algoritmos para ajustar modelos multivariantes, tanto exploratorios como de regresión, en presencia de datos faltantes. Estos algoritmos sirven además como estrategias de preprocesado de los datos antes del uso de cualquier otro método. Respecto a las aplicaciones, en este bloque se explora la reconstrucción de redes en ciencias ómicas cuando aparecen valores faltantes o atípicos en las bases de datos. Una segunda aplicación de esta parte es la transferencia de modelos de calibración entre instrumentos de infrarrojo cercano, evitando así costosas re-calibraciones en bioindustrias y laboratorios de investigación. Finalmente, se propone un paquete software que incluye una interfaz amigable, disponible de forma gratuita para imputación de datos faltantes. En la última parte, se discuten los aspectos más relevantes de esta tesis para la investigación y la biotecnología, incluyendo líneas futuras de trabajo.
Aquesta tesi doctoral es centra en l'estudi, desenvolupament, i aplicació de tècniques quimiomètriques en l'emergent camp de la biologia de sistemes. Procediments comúnment utilizats i mètodes nous s'apliquen per a resoldre preguntes d'investigació en diferents equips multidisciplinars, tant en l'àmbit acadèmic com en l'industrial. Les metodologies desenvolupades en aquest document enriquixen la plétora de tècniques utilitzades en les ciències òmiques per a entendre el funcionament d'organismes biològics i milloren els processos en la indústria biotecnològica, integrant coneixement biològic a distints nivells i explotant els paquets de software derivats d'aquesta tesi. Aquesta dissertació s'estructura en quatre parts. El primer bloc descriu el marc en el qual s'articulen les contribucions ací presentades. En ell s'esbossen els objectius dels dos projectes d'investigació relacionats amb aquesta tesi. Així mateix, s'introduixen els temes específics desenvolupats en aquest document mitjançant presentacions en conferències i articles d'investigació. En aquesta part figura una descripació exhaustiva de les ciències òmiques i les seues interrelacions en el paradigma de la biologia de sistemes, junt amb una revisió dels mètodes multivariants més aplicats en quimiometria, que supossen els pilars sobre els quals s'assenten els nous procediments ací proposats. La segona part es centra en resoldre problemes dins de la metabolòmica, fluxòmica, proteòmica i genòmica a partir de l'anàlisi de dades. Per a això es proposen diverses alternatives per a compendre a grans trets les dades de fluxos metabòlics en estat estacionari. Algunes d'elles estàn basades en l'aplicació de mètodes multivariants propostos amb anterioritat, mentre que altres són tècniques noves basades en descomposicions bilineals utilizant rutes metabòliques elementals. A partir d'aquestes s'ha desenvolupat software de lliure accés per a la comunitat científica. Al seu torn, en aquesta tesi es proposa un marc per a analitzar dades metabòliques en estat no estacionari. Per a això s'adapta l'enfocament tradicional per a sistemes en estat estacionari, modelant les dinàmiques dels experiments utilizant anàlisi de dades de dues i tres vies. En aquesta part de la tesi també s'establixen relacions entre els distints nivells òmics, integrant diferents fonts d'informació en models de fusió de dades. Finalment, s'estudia la interacció entre organismes, com taronges i fongs, mitjançant l'anàlisi multivariant d'imatges, amb futures aplicacions a la indústria alimentària. El tercer bloc d'aquesta tesi representa un estudi a fons de diferents problemes relacionats amb dades faltants en quimiometria, biologia de sistemes i en la indústria de bioprocessos. En els capítols més teòrics d'aquesta part, es proposen nous algoritmes per a ajustar models multivariants, tant exploratoris com de regressió, en presencia de dades faltants. Aquests algoritmes servixen ademés com a estratègies de preprocessat de dades abans de l'ús de qualsevol altre mètode. Respecte a les aplicacions, en aquest bloc s'explora la reconstrucció de xarxes en ciències òmiques quan apareixen valors faltants o atípics en les bases de dades. Una segona aplicació d'aquesta part es la transferència de models de calibració entre instruments d'infrarroig proper, evitant així costoses re-calibracions en bioindústries i laboratoris d'investigació. Finalment, es proposa un paquet software que inclou una interfície amigable, disponible de forma gratuïta per a imputació de dades faltants. En l'última part, es discutixen els aspectes més rellevants d'aquesta tesi per a la investigació i la biotecnologia, incloent línies futures de treball.
Folch Fortuny, A. (2016). Chemometric Approaches for Systems Biology [Tesis doctoral no publicada]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/77148
TESIS
Premiado

Activation of cytotoxic T cells in human requires specific binding of antigenic peptides to human leukocyte antigen (HLA) molecules. HLA is the most polymorphic protein in the human body, currently 1814 different alleles collected in the HLA sequence database at the European Bioinformatics Institute. Most of the HLA molecules recognise different peptides. Also, some peptides can be recognised by several of HLA molecules. In the present project, all available class I HLA alleles are classified into supertypes. Super - binding motifs for peptides binding to some supertypes are defined where binding data are available. A variety of chemometric techniques are used in the project, including 2D and 3D QSAR techniques and different variable selection methods like SIMCA, GOLPE and genetic algorithm. Principal component analysis combined with molecular interaction fields calculation by the program GRID is used in the class I HLA classification. This thesis defines an HLA-A3 supermotif using two QSAR methods: the 3D-QSAR method CoMSIA, and a recently developed 2D-QSAR method, which is named the additive method. Four alleles with high phenotype frequency were included in the study: HLA-A*0301, HLA-A*1101, HLA-A*3101 and HLA- A*6801. An A*020T binding motif is also defined using amino acid descriptors and variable selection methods. Novel peptides have been designed according to the motifs and the binding affinity is tested experimentally. The results of the additive method are used in the online server, MHCPred, to predict binding affinity of unknown peptides. In HLA classification, the HLA-A, B and C molecules are classified into supertypes separately. A total of eight supertypes are observed for class I HLA, including A2, A3, A24, B7, B27, B44, CI and C4 supertype. Using the HLA classification, any newly discovered class I HLA molecule can be grouped into a supertype easily, thus simplifying the experimental function characterisation process.

Throughout the hydrocarbon supply chain, process optimisation is driven by the desire to maximise profit margins. In the global refining marketplace, the biggest cost is crude oil and to improve margins increasing use of non-conventional crude oils (also called opportunity crudes) lowers the cost of the crude blend. Opportunity crudes are selected based on market forces, for example in North America, the production booms in shale oil and tar sands have provided ample amounts of new low-cost oils which refineries are buying and processing. However, as these oils are new to the marketplace many refineries have never processed them before which brings about challenges. These are mainly a lack of understanding of the quality of the crude oil being processed (shale oils for example can come from many thousands of wells) and how these oils interact with the more conventional refinery feedstocks (such as Brent or West Texas Intermediate). The Eng.D project was carried out in collaboration with Intertek Group plc, a multinational corporate organisation consisting of more than 42,000 employees in over 1,000 locations in over 100 countries across the globe, and was aimed at developing solutions to address crude oil processing problems. The issues covered over the course of the project fall into the areas of: enhancing understanding of crude oil quality, addressing issues of hydrocarbon blend stability because of blending and better utilisation of process data to promote efficiency and facilitate process troubleshooting. As such, the Eng.D project was firstly concerned with developing a robust chemometric model, based on Near Infrared spectra, for use in a major Asian refinery. Once built and tuned this model was ultimately used to predict physical properties (such as density, sulphur content and distillation properties) of every crude oil delivery and also online in the refinery for frequent prediction of crude oil blend properties. The second project was then aimed at solving refinery issues of the deposition of undesirable material (such as wax and asphaltenes) in pipes and process units. The research carried out during the course of the Eng.D project resulted in a patented approach to characterise these issues and provide refineries strategies to mitigate the problems. This approach is not just limited to crude oils but can be applied to any blended hydrocarbon streams and detects the precipitation of undesirable material using Near Infrared spectroscopy and microscopy. This ii approach has now been applied to solving problems of blending crude oils in refineries and offshore, heavy fuel oils, shale oils and marine fuels. Finally, the application of smart data analytics in an upstream installation was investigated. The objective of this application was to provide a customer with process troubleshooting for a historical recurring pump failure issue. To achieve this, the root cause of the issue first needed to be identified and then a solution developed.

De nombreux systèmes biologiques sont intrinsèquement polydispersés, présentant de multiples espèces coexistantes, de taille, de forme ou de conformation différentes (c'est-à-dire, mélanges oligomèriques, des complexes faiblement liés se dissociant en composantes individuelles ou des espèces apparaissant lors de processus amyloïdogéniques). L'étude de tels systèmes complexes est une tâche difficile en raison de l'instabilité des espèces concernées, de leurs concentrations relatives faibles et interdépendantes et des difficultés rencontrées pour l'isolation des composantes pures. Dans cette thèse, j'ai développé des approches méthodologiques pour appliquer la diffusion des rayons X aux petits angles (SAXS), une technique de biologie structurale, à l'étude de systèmes polydispersés. SAXS est une technique additive et par conséquent, le diagramme de diffusion mesuré pour un échantillon polydispersé correspond à la somme pondérée en concentration des contributions de chacune des composantes individuelles du mélange. Cependant, la décomposition des données de SAXS en des spectres spécifiques des espèces et de leurs concentrations relatives est extrêmement laborieuse et ambigue. Dans cette thèse, je présente d'abord une approche objective pour solidement décomposer les jeux de données de SAXS en composantes individuelles. Cette approche adapte la méthode chimiométrique « Multivariable Curve Resolution Alternate Least Squares » (MCR-ALS) aux spécificités des données de SAXS. Notre méthode permet une décomposition rigoureuse et robuste des données de SAXS en introduisant simultanément différentes représentations de ces données et par conséquent, en mettant l'accent sur des changements moléculaires à différentes plages de temps et de résolution structurale. Nous avons appliqué cette approche, que nous appelons COSMiCS (Analyse structurelle objective complexe des systèmes multi-composants) pour étudier deux systèmes polydispersés: la fibrillation des protéines, et les fluctuations conformationnelles de protéines grâce à l'analyse de données obtenues à l'aide d’une technique de couplage de chromatographie d'exclusion de taille (SEC) avec le ligne de SAXS (SEC-SAXS). L'importance d'étudier les processus de fibrillation réside dans leur implication dans des pathologies amyloïdogéniques telles que les maladies de Parkinson ou d'Alzheimer. Il existe de fortes indications que les espèces oligomériques solubles, et non les fibrilles matures, sont la cause principale de la cytotoxicité et des dommages neuronaux. Cette observation souligne l'importance de caractériser les premiers stades des processus de fibrillation. Notre approche COSMiCS a permis d'étudier les processus amyloïdogéniques de l'insuline et du mutant familial E46K de l'α-synucléine, une protéine associée à la maladie de Parkinson. Cette analyse permet la caractérisation structurale des espèces présentes (y compris les espèces oligomériques) et la caractérisation cinétique de leurs transformations.La deuxième partie de la thèse est consacrée à l'utilisation de COSMiCS pour analyser des données de SEC-SAXS. Le SEC-SAXS est extrêmement populaire et a été implémenté sur plusieurs lignes de SAXS à travers le monde. En utilisant des données synthétiques, je démontre la capacité des approches chimiométriques à décomposer des profils chromatographiques complexes. À l'aide de cette approche, j'ai décomposé l’ensemble des données SEC-SAXS mesurés pour la Prolyl OligoPeptidase (POP).En résumé, cette thèse présente une nouvelle approche chimiométrique qui peut être généralement appliquée à tout mélange macromoléculaire pouvant subir une modifacation de son équilibre et pouvant être abordé par SAXS. Les complexes biomoleculaires transitoires, les processus de repliement, les réarrangements structuraux dépendants d’un ligand ou la formation de grands ensembles supramoleculaires peuvent être sondés de façon structurale en utilisant l'approche COSMiCS
Many biological systems are inherently polydisperse, presenting multiple coexisting species differing in size, shape or conformation (i.e. oligomeric mixtures, weakly bound complexes, and species appearing along amyloidogenic processes). The study of such complex systems is challenging due to the instability of the species involved, their low and interdependent relative concentrations, and the difficulties to isolate the pure components. In this thesis, I have developed methodological approaches to apply Small-Angle X-ray Scattering (SAXS), a low-resolution structural biology technique, to the study of polydisperse systems. As an additive technique, the SAXS pattern measured for a polydisperse sample corresponds to the concentration-weighted sum of the contributions from each of the individual components. However, decomposition of SAXS data into species-specific spectra and relative concentrations is laborious and burdened by ambiguity. In this thesis, I present an approach to decompose SAXS datasets into the individual components. This approach adapts the chemometrics Multivariate Curve Resolution Alternating Least Squares (MCR-ALS) method to the specificities of SAXS data. Our method enables the rigorous and robust decomposition of SAXS data by simultaneously introducing different representations of these data and, consequently, emphasizing molecular changes at different time and structural resolution ranges. We have applied this approach, which we name COSMiCS (Complex Objective Structural analysis of Multi-Component Systems), to study two polydisperse systems: amyloid fibrillation by analysing time-dependent SAXSdata, and conformational fluctuations through the analysis of data obtained using on-line size-exclusion chromatography coupled to SAXS (SEC-SAXS). The importance of studying fibrillation processes lies in their implication in amyloidogenic pathologies such as Parkinson’s or Alzheimer’s diseases. There exist strong indications that soluble oligomeric species, and not mature fibrils, are the main cause of cytotoxicity and neuronal damage emphasizing the importance of characterizing early stages of fibrillation. The first application of our COSMiCS approach has allowed the study of the amyloidogenic mechanisms of insulin and the familial mutant E46K of ↵-synuclein, a Parkinson’s disease related protein. The analysis enables the structural characterization of all the species present as well as their kinetic transformations. The second part of the thesis is dedicated to the use of COSMiCS to analyze on-line SEC-SAXS experiments. Using synthetic data, I demonstrate the capacity of chemometric approaches to decompose complex chromatographic profiles. Using this approach, I have studied the conformational fluctuations in prolyl oligopeptidase (POP), a protein related to synaptic functions and neuronal development. In summary, this thesis presents a novel chemometrics approach that can be generally applied to any macromolecular mixture with a tuneable equilibrium that is amenableto SAXS. Transient biomolecular complexes, folding processes, or ligand-dependent structural rearrangements can be probed structurally using COSMiCS

Contrairement à l’analyse ciblée des composés volatils du vin par chromatographie en phase gazeuse couplée à la spectrométrie de masse (GC-MS), les approches par GC-MS non ciblées prennent en compte les composés connus et inconnus. Ces méthodes sont plus rapides et fournissent une représentation plus complète de la composition de l’échantillon. Bien que plusieurs approches non-ciblées aient été développées, il y a encore une forte demande d’outils automatisés pour le traitement des données, en particulier pour les données multidimensionnelles complexes telles que celles de multiples chromatogrammes GC-MS. Ce travail visait à développer deux nouvelles approches chimiométriques pour l’analyse des données GC-MS non ciblées. Ces approches prennent en considération les décalages de temps de rétention entre les échantillons et rendent inutile l’intégration des pics. Elles ont été testées avec un jeu de données GC-MS simulées et un jeu de données GC-MS réelles d’échantillons de vin. De plus, l’une des deux approches GC-MS non ciblée a été combinée à la technique d’analyse sensorielle rapide de "projective mapping". Cette méthodologie a été utilisée pour étudier l’impact de la fermentation malolactique sur des vins issus du cépage Pinotage ainsi que l’effet de l’âge de la vigne, de la turbidité du moût et de la souche de levure sur l’arôme de vins de Riesling expérimentaux
In contrast to targeted gas chromatography mass spectrometry (GC-MS) analysis of wine volatiles, non-targeted GC-MS approaches take information of known and unknown compounds into account, are faster, inherently more comprehensive and give a more holistic representation of the sample composition. Although several non-targeted approaches have been developed, there is still a great demand for automated data processing tools, especially for complex multi-way data such as chromatographic data obtained from multichannel detectors (e.g. GC-MS chromatograms of multiple samples). This work therefore aimed at the development of data processing procedures for non-targeted GC-MS analysis of volatile wine compounds. The two developed approaches use basic matrix manipulation of segmented GC-MS chromatograms and PCA or PARAFAC multi-way modelling. The approaches take retention time shifts between samples into account and avoid peak integration. A demonstration of the new fingerprinting approaches is presented using an artificial GC-MS data set and an experimental full-scan GC-MS data set obtained for a set of experimental wines. Results of the new approaches were also compared to a references method. Furthermore, the combination of one of the developed GC-MS fingerprinting approaches with the fast sensory screening technique projective mapping was exploited as a powerful approach to simultaneously study the volatile composition and the sensory characteristics of experimental wines. This methodology was used to study the impact of different malolactic fermentation scenarios on two different Pinotage wine styles and for a full factorial investigation of the impact of grape vine age, must turbidity and yeast strain on the aroma of Riesling experimental wines

La metabolomica, che consiste nella identificazione di tutti i metaboliti presenti all’interno dei campioni biologici analizzati, è un approccio ampiamente applicato in diversi campi di ricerca quali: identificazione di biomarcatori, sviluppo di nuovi farmaci, scienze alimentari e ambientali. La metabolomica è strettamente legata alla capacità di tecniche analitiche fra queste una delle più applicate è la gas cromatografia accoppiata alla spettrometria di massa. Moderne piattaforme analitiche possono generare centinaia di migliaia di spettri, rilevando una quantità impressionante di molecole distinte. Nonostante i progressi tecnici raggiunti sul lato sperimentale, la conversione dei segnali misurati dagli strumenti in informazioni utili non è un passaggio scontato in studi metabolomici. Per ogni composto identificato, l’obbiettivo è ottenere la concentrazione relativa tra tutti i campioni analizzati e lo spettro di massa associato al composto, necessario per l’identificazione della molecola stessa. I software disponibili per l’analisi dei dati sperimentali sono stati ripetutamente indicati come una fonte importante di incertezza, limitando fortemente sia la quantità che la qualità delle informazioni estratte. Gli strumenti più applicati richiedono l’impostazione di diversi parametri da parte dell’operatore, influenzando il risultato dell’analisi. In questa tesi è descritto un nuovo approccio, chiamato AutoDise, per l’analisi dei dati GC-MS. L’elaborazione dei segnali sperimentali si basa su PARAFAC2. PARAFAC2 è un modello che scompone dati multidimensionali, discriminando tra i diversi segnali nei campioni. Grazie alle sue proprietà, PARAFAC2 non ha bisogno che i dati siano pretrattati e non richiede di impostare parametri, mentre software utilizzati in questo ambito richiedono di definire diversi parametri e un laborioso pretrattamento dei dati, richiedendo l’intervento di un utente esperto, inoltre la riproducibilità dei risultati è limitata, dipendendo i parametri scelti dall’utente. Tuttavia, il fitting di modelli PARAFAC2 coinvolge diverse fasi ed è necessario un esperto analista per l’analisi e l’interpretazione dei modelli. AutoDise è un sistema esperto in grado di gestire tutti i passaggi riguardanti la modellazione e di generare una tabella dei picchi in cui ogni composto è identificato in modo univoco, con risultati completamente riproducibili. Questo è possibile grazie alla combinazione di diversi strumenti diagnostici e grazie all’ applicazione di modelli d’intelligenza artificiale. Le prestazioni dell’approccio sono state testate su un complesso dataset di oli d’oliva ottenuto tramite analisi GC-MS. I dati sono stati analizzati sia manualmente, da utenti esperti, sia automaticamente con il metodo AutoDise proposto e le tabelle dei picchi risultanti sono state confrontate. I risultati mostrano che AutoDise supera l’analisi manuale sia in termini di numero di composti identificati che di qualità dell’identificazione e della quantificazione. Inoltre, è stata sviluppata una GUI per rendere l’algoritmo più accessibile a persone non esperte nel linguaggio di programmazione. La tesi include un tutorial che mostra le caratteristiche principali e come utilizzare l’interfaccia grafica. Un’altra parte importante della tesi è stata dedicata al test e allo sviluppo di nuove reti neurali artificiali da implementare nel software AutoDise per rilevare quali componenti PARAFAC2 stanno fornendo informazioni chimicamente utili. A tal fine, più di 170.000 profili sono stati etichettati manualmente, al fine di addestrare, validare e testare una rete neurale convoluzionale e una rete bilineare con memoria a breve termine e un modello k-nearest neighbour. I risultati suggeriscono che le reti di deep learning possono essere efficacemente applicate per la classificazione automatica dei profili cromatografici.
Metabolomics, which consists of identifying all the metabolites present in the biological samples analysed, is an approach widely applied in various research fields such as biomarker identification, new drug development, food and environmental sciences. Metabolomics is closely linked to the ability of analytical techniques, one of the most widely applied being gas chromatography coupled to mass spectrometry. Modern analytical platforms can generate hundreds of thousands of spectra, detecting an impressive number of distinct molecules. Despite the technical progress achieved on the experimental side, the conversion of signals measured by instruments into useful information is not an obvious step in metabolomic studies. For each identified compound, the goal is to obtain the relative concentration among all analysed samples and the mass spectrum associated with the compound needed to identify the molecule itself. The software available for analysing experimental data has repeatedly been cited as a major source of uncertainty, severely limiting both the quantity and quality of the information extracted. The most applied tools are based on univariate data analysis, considering each sample separately from the others and requiring the operator to set several parameters, affecting the result of the analysis. In this thesis, a new approach, called AutoDise, for the analysis of GC-MS data is described. The processing of the experimental signals is based on PARAFAC2. PARAFAC2 is a model that decomposes multidimensional data, discriminating between different signals in the samples. Due to its properties, PARAFAC2 does not need the data to be pre-processed and does not require parameters to be set, whereas software used in this field requires several parameters to be defined and laborious pre-processing of the data, requiring the intervention of an expert user, and the reproducibility of the results is limited, depending on the parameters chosen by the user. However, fitting PARAFAC2 models involves several steps and an experienced analyst is needed to analyse and interpret the models. AutoDise is an expert system capable of handling all modelling steps and generating a peak table in which each compound is uniquely identified, with fully reproducible results. This is possible thanks to the combination of different diagnostic tools and the application of artificial intelligence models. The performance of the approach was tested on a complex dataset of olive oils obtained by GC-MS analysis. The data were analysed both manually, by experienced users, and automatically with the proposed AutoDise method and the resulting peak tables were compared. The results show that AutoDise outperforms manual analysis both in terms of the number of compounds identified and the quality of identification and quantification. In addition, a GUI was developed to make the algorithm more accessible to people not skilled in the programming language. The thesis includes a tutorial showing the main features and how to use the GUI. Another important part of the thesis was devoted to testing and developing new artificial neural networks to be implemented in the AutoDise software to detect which PARAFAC2 components are providing chemically useful information. To this end, more than 170,000 profiles were manually labelled in order to train, validate and test a convolutional neural network and a bilinear network with short-term memory and a k-nearest neighbour model. The results suggest that deep learning networks can be effectively applied for the automatic classification of chromatographic profiles.

In the recent years, the consumer choices have been focused on health-promoting plant-based food and their preferences are oriented towards regional foodstuff from local productions. Therefore, an important factor for vegetables grown Trentino-Alto Adige (Italy) is to point out the added value of alpine farming to evaluate the nutritional values of farming products. Omics technologies (e.g. genomics, transcriptomics, proteomics and metabolomics) are aimed at investigating the assessment of different pools of molecules and how they are translated into the structure, function, and dynamics of a biological system or systems in order to provide a comprehensive characterization of a specific organism. Research use the omics techniques to exhaustively understand the functionality of food components. Several sophisticated chromatographic methods, spectroscopic techniques and chemometric tools are applied to give an insight into a comprehensive overview of the intrinsic quality, typicality and regionality of specific plant-based foods in the present PhD thesis: apples and potatoes. The quality of these foods is evaluated by quantifying the secondary metabolites to investigate their nutraceutical values. The aim of this PhD project is to use several analytical techniques (LC-MS, UV-VIS) that are capable of comprehensively characterizing the food metabolome with particular emphasis on those components with high nutritional values. The data analysis and data handling of omics data requires advanced bioinformatic, statistical, and chemometric tools. Potatoes and apples are chosen as target matrices for these studies for their relevance in the local economy and for the peculiar chemical composition of particular interest for their health-promoting proprieties. The information is acquired using several sophisticated chromatographic and spectroscopic techniques, such as ultra-high performance liquid chromatography coupled with triple quadrupole mass spectrometry (UHPLC– MS/MS) and UV/VIS. It is integrated to chemometric approaches (principal component analysis (PCA), partial least square regression (PLS), and data fusion) to achieve a comprehensive targeted chemical characterization. The sampling procedures gathers, in the case of the potatoes study, reference cultivars that may be found in the common retailers of Trentino/Alto-Adige and different production areas, the apples of 22 cultivars were harvest from the fields of the Laimburg Research Centre (Vadena, Italy) to guaranty comparability of the obtained data. Our results may be used as solid foundation for a reliable evaluation of apples and potatoes healthy "potential" value based on cutting-edge techniques, which are capable of providing comprehensive data regarding the alpine food quality parameters with high efficiency and reliability

The WHO classification of brain tumors is based on histological features and the aggressiveness of the tumor is classified from grade I to IV, where grade IV is the most aggressive. Today, the correlation between prognosis and tumor grade is the most important component in tumor classification. High grade gliomas, glioblastomas, are associated with poor prognosis and a median survival of 14 months including all available treatments. Low grade meningiomas, usually benign grade I tumors, are in most cases cured by surgical resection. However despite their benign appearance grade I meningiomas can, without any histopathological signs, in some cases develop bone invasive growth and become lethal. Thus, it is necessary to improve conventional treatment modalities, develop new treatment strategies and improve the knowledge regarding the basic pathophysiology in the classification and treatment of brain tumors. In this thesis, both proteomics and metabolomics have been applied in the search for biomarkers or biomarker patterns in two different types of brain tumors, gliomas and meningiomas. Proteomic studies were carried out mainly by surface enhanced laser desorption ionization time of flight mass spectrometry (SELDI-TOF-MS). In one of the studies, isobaric tags for relative and absolute quantitation (iTRAQ) labeling in combination with high-performance liquid chromatography (HPLC) was used for protein detection and identification. For metabolomics, gas-chromatography time-of-flight mass spectrometry (GC-TOF-MS) has been the main platform used throughout this work for generation of robust global metabolite profiles in tissue, blood and cell cultures. To deal with the complexity of the generated data, and to be able to extract relevant biomarker patters or latent biomarkers, for interpretation, prediction and prognosis, bioinformatic strategies based on chemometrics were applied throughout the studies of the thesis. In summary, we detected differentiating protein profiles between invasive and non-invasive meningiomas, in both fibrous and meningothelial tumors. Furthermore, in a different study we discovered treatment induce protein pattern changes in a rat glioma model treated with an angiogenesis inhibitor. We identified a cluster of proteins linked to angiogenesis. One of those proteins, HSP90, was found elevated in relation to treatment in tumors, following ELISA validation. An interesting observation in a separate study was that it was possible to detect metabolite pattern changes in the serum metabolome, as an effect of treatment with radiotherapy, and that these pattern changes differed between different patients, highlighting a possibility for monitoring individual treatment response.  In the fourth study of this work, we investigated tissue and serum from glioma patients that revealed differences in the metabolome between glioblastoma and oligodendroglioma, as well as between oligodendroglioma grade II and grade III. In addition, we discovered metabolite patterns associated to survival in both glioblastoma and oligodendroglioma. In our final work, we identified metabolite pattern differences between cell lines from a subgroup of glioblastomas lacking argininosuccinate synthetase (ASS1) expression, (ASS1 negative glioblastomas), making them auxotrophic for arginine, a metabolite required for tumor growth and proliferation, as compared to glioblastomas with normal ASS1 expression (ASS1 positive). From the identified metabolite pattern differences we could verify the hypothesized alterations in the arginine biosynthetic pathway. We also identified additional interesting metabolites that may provide clues for future diagnostics and treatments. Finally, we were able to verify the specific treatment effect of ASS1 negative cells by means of arginine deprivation on a metabolic level.

The use of online spectroscopic analysis to predict final properties of in-process pulp has the capacity to revolutionize the paper industry saving time and money in process control and quality assurance. Water has a very low Raman scattering cross section, which makes this spectroscopic probe ideal for classification typical pulps with high and varying water content. Work described in this thesis aims to develop new methods to predict pulp end product properties from the Raman spectra of in process pulps. We predict breaking length of wet and dry pulps from their Raman spectra using a laboratory Raman probe. Pulp samples were refined to five levels of refining energy by Canfor Pulp Ltd. We determined that to accurately predict breaking length, pulp spectra must first, before modelling, be sorted into groups based on refining energy. Dry pulps yield better prediction models then wet pulps. Breaking length can be determined within 10% error after discrete wavelet transform (DWT) and template oriented genetic algorithm (TOGA) preprocessing, however, industry standard is 5%. Principal component analysis (PCA) suggests that Raman measurements made on the side edge of a pulp sheet are similar to those made on the top surface. We designed and built a Raman probe set-up for the investigation of industry-standard brightness pads in a modified PulpEye sample chamber to mimic online monitoring. We fashioned a 5-axis stage, from optical positioning components, to hold an INNO spectrometer for easy alignment through a small sapphire widow in the sample holder. Spectra acquired with the INNO-PulpEye set-up predict breaking length with an accuracy similar to the spectra taken with the laboratory set-up. Utilizing DWT for background suppression and noise reduction as well as TOGA for feature selection we were able to predict breaking lengths with 6-10% error using both the lab and PulpEye sample chamber Raman probe systems. We have also investigated the possibility of predicting pulp viscosity from its Raman spectrum. We were able to predict viscosity over the full range with errors below 6% and 1% for pulps with viscosities between 500-560 mL/g.
Science, Faculty of
Chemistry, Department of
Graduate

Questo progetto di dottorato studia le differenti applicazioni delle tecniche ottiche non distruttive per la valutazione della qualità e della shelf-life di prodotti vegetali così come l’identificazione precoce di sviluppi microbici su superfici industriali. La spettroscopia, l’analisi dell’immagine e l’analisi dell’immagine iperspettrale possono giocare un ruolo importante nella valutazione sia della qualità che della sicurezza degli alimenti grazie alla rapidità e sensibilità della tecnica, specialmente quando si utilizzano strumenti semplificati portatili. Un approccio statistico multivariato (chemiometria) è richiesto al fine di estrarre informazioni dal segnale acquisito, riducendo la dimensionalità dei dati e mantenendo le informazioni spettrali più utili. Lo scopo del primo studio presentato – Testing of a Vis-NIR system for the monitoring of long-term apple storage – è la valutazione dell’applicabilità della spettroscopia nel visibile e vicino infrarosso (Vis-NIR) per il monitoraggio e la gestione delle mele durante lo stoccaggio a basse temperature. Per sette mesi è stata seguita l’evoluzione in termini di grado zuccherino e consistenza delle mele suddivise in classi di maturazione. I risultati hanno indicato che la spettroscopia è una tecnica non-distruttiva che consente una stima accurata dei parametri chimico-fisici per la classificazione delle mele in lotti omogenei. Il lavoro descritto nel secondo paragrafo - Wavelength selection with a view to a simplified handheld optical system to estimate grape ripeness – è finalizzato all’identificazione delle tre lunghezze d’onda più importanti per il riconoscimento, direttamente in campo, dell’uva pronta per essere raccolta al fine della messa a punto di un sistema semplificato e a basso costo. I coefficienti di regressione standardizzati del modello PLS (Partial Least Square) sono stati utilizzati per selezionare le variabili più importanti, che racchiudono l’informazione più utile lungo l’intero spettro. La stessa procedura è stata condotta per determinare la freschezza delle foglie di Valerianella durante la shelf-life - Selection of optimal wavelengths for decay detection in fresh-cut Valerianella Locusta laterr (terzo paragrafo). Lo scopo del lavoro presentato nel quarto paragrafo del primo capitolo - Comparison between FT-NIR and Micro-NIR in the evaluation of Acerola fruit quality, using PLS and SVM regression algorithms – è stimare l’acidità titolabile e il contenuto di acido ascorbico all’interno del frutto acerola, utilizzando uno strumento compatto e a basso costo denominato Micro-NIR, che lavora nell’intervallo spettrale 950-1650 nm. I dati spettrali sono stati modellati mediante l’applicazione di due algoritmi PLS e SVM (Support Vector Machine). La capacità predittiva dello strumento semplificato è risultata interessante per applicazioni di monitoraggio in campo, soprattutto modellizzando i dati in modo non lineare. Nel secondo capitolo, è presentata l’applicazione di immagini RGB per la valutazione delle superfici - Image texture analysis, a non-conventional technique for early detection of biofilm. La texture dell’immagine è definita come una differenza nella distribuzione spaziale, nella frequenza e nell’intensità dei livelli di grigio in ogni pixel dell’immagine. Questo metodo è stato determinante per l’identificazione precoce dello sviluppo microbico su superfici normalmente impiegate nell’industria alimentare. L’approccio chemiometrico è stato cruciale in ogni fase del progetto di dottorato ed è definito come un approccio statistico multivariato che si applica ai dati chimici per estrarre informazione utile, ridurre il rumore di fondo e l’informazione ridondante. Il lavoro presentato all’inizio del terzo capitolo - Hyperspectral image analysis: a tutorial - propone una procedura standard per l’elaborazione di dati tridimensionali, presentando un esempio relativo alla predizione del raffermamento del pane in cassetta. Il secondo paragrafo del terzo capitolo, presenta una applicazione dell’immagine iperspettrale su acerola, focalizzata sul contenuto di vitamina C - HSI for quality evaluation of vitamin C content in Acerola fruit. In questo lavoro, è stata acquisita l’immagine di dieci acerola, raccolte in funzione del livello di maturazione, definito in base al colore della buccia (cinque acerola verdi e cinque rosse). Lo spettro della polvere di vitamina C pura è stato utilizzato come riferimento per l’applicazione di due algoritmi di correlazione (spectral angle mapping e correlation coefficient), consentendo la costruzione di mappe qualitative di distribuzione dell’acido ascorbico all’interno del frutto. Lo scopo dell’ultimo lavoro presentato è la valutazione della qualità post raccolta dell’acerola - Selection of NIR wavelengths from hyperspectral imaging data for the quality evaluation of Acerola fruit. Le immagini iperspettrali di venti acerola sono state acquisite per cinque giorni consecutivi. La valutazione delle modificazioni spettrali durante il tempo ha consentito la selezione delle tre lunghezze d’onda caratterizzanti il processo di maturazione/degradazione del frutto. L’immagine in falsi colori, derivante dalla composizioni delle immagini alle tre lunghezze d’onda di interesse, consente l’identificazione precoce del processo degradativo in maniera rapida e non distruttiva. Le tre tecniche non distruttive impiegate in questo progetto di dottorato hanno dimostrato efficienza e applicabilità per la valutazione della qualità e della sicurezza degli alimenti, rispondendo alla necessità dell’industria alimentare di tecniche accurate, veloci e obiettive per assicurare produzioni ottimali lungo l’intero processo produttivo.
This PhD project regards different applications of non-destructive optical techniques to evaluate quality and shelf life of agro-food product as well as the early detection of biofilm on food plants. Spectroscopy, image analysis and hyperspectral imaging could play an important role in the assessment of both quality and safety of foods due to their rapidity and sensitivity especially when using simplified portable devices. Due to the huge amount of collected data, chemometric, a multivariate statistical approach, is required, in order to extract information from the acquired signals, reducing dimensionality of the data while retaining the most useful spectral information. The thesis is organized in four chapters, one for each technique and a final chapter including the overall conclusion. Each chapter is divided in case studies according to the matrix analysed and the data acquisition and elaboration carried out. The first chapter is about spectroscopy. The aim of the first study - Testing of a Vis-NIR system for the monitoring of long-term apple storage - is to evaluate the applicability of visible and near-infrared (Vis-NIR) spectroscopy to monitor and manage apples during long-term storage in a cold room. The evolution of the apple classes, originally created, was analysed during 7 months of storage by monitoring TSS and firmness. Vis-NIR allows an accurate estimation of chemical-physical parameters of apples allowing a non-destructive classification of apples in homogeneous lots and a better storage management. The work reported in the second paragraph - Wavelength selection with a view to a simplified handheld optical system to estimate grape ripeness - is aimed to identify the three most significant wavelengths able to discriminate grapes ready to be harvested directly in the field. Wavelengths selection was carried out with a view to construct a simplified handheld and low-cost optical device. Standardized regression coefficients of the PLS model were used to select the relevant variables, representing the most useful information of the full spectral region. The same approach was followed to discriminate freshness levels during shelf-life of fresh-cut Valerianella leaves - Selection of optimal wavelengths for decay detection in fresh-cut Valerianella Locusta Laterr. (third paragraph). The aim of the work presented in the fourth paragraph of the first chapter - Comparison between FT-NIR and Micro-NIR in the evaluation of Acerola fruit quality, using PLS and SVM regression algorithms - is to estimate titratable acidity and ascorbic acid content in acerola fruit, using a MicroNIR, an ultra-compact and low-cost device working between 950 – 1650 nm. The spectral data were modelled using two different regression algorithms, PLS (partial least square) and SVM (support vector machine). The prediction ability of Micro-NIR appears to be suitable for on field monitoring using non-linear regression modelling (i.e. SVM). In the second chapter, image analysis was performed. The traditional RGB imaging for the evaluation of image texture, a specific surface characteristic, is presented. The texture of an image is given by differences in the spatial distribution, in the frequency and in the intensity of the values of the grey levels of each pixel of the image. This technique was applied for the early detection of biofilm in its early stages of development, when it is still difficult to observe it by the naked eye, was evaluated (Image texture analysis, a non-conventional technique for early detection of biofilm). In the third paragraph, image and spectroscopy were combined in hyperspectral imaging applications. Data analysis by chemometric was crucial in any stage of my PhD project. Chemometric is a multivariate statistical approach that is applied on chemical data to extract the useful information avoiding noise and redundant data. At the beginning of the third chapter - Hyperspectral image analysis: a tutorial - proposes an original approach, developed as a flow sheet for three-dimensional data elaboration. The method was applied, as an example, to the prediction of bread staling during storage. The first application about hyperspectral on acerola is focused on the vitamin C content - HIS for quality evaluation of vitamin C content in Acerola fruit. Ten different acerola fruits picked up according to two different stages of maturity, based on the colour of the peel (5 green and 5 red acerola), were analysed. The spectra of pure vitamin C powder was used as references for computing models with two different correlation techniques: spectral angle mapping and correlation coefficient allowing the construction of a qualitative distribution map of ascorbic acid inside the fruit. The aim of the last one work presented is to evaluate acerola post-harvest quality - Selection of NIR wavelengths from hyperspectral imaging data for the quality evaluation of Acerola fruit. Hyperspectral images of 20 acerolas were acquired for five consecutive days and an investigation of time trends was carried out to highlight the most important three wavelengths that characterized the ripeness/degradation process of the Acerola fruit. The false-colour RGB images, derived from the composition of the three interesting wavelengths selected, data enable early detection of the senescence process in a rapid and non-destructive manner. In conclusion, the three non-destructive optical techniques applied in this PhD project have proved to be one of the most efficient and advanced tools for safety and quality evaluation in food industry answering the need for accurate, fast and objective food inspection methods to ensure safe production throughout the entire production process.

In this Ph.D. project it has been developed a study of multicomponent chromatograms of complex mixtures, using a chemometric approach. The activity has been concentrated in the study of analytical-separative methods (in particular Gas Cromatography-Mass Spectrometry, GC-MS) for complex samples of environmental interest, especially for PM (particulate matter) samples. A fundamental part of this Ph.D. project has been dedicated to the development of mathematical and statistical algorithms for the data treatment of the GC-MS signal obtained from the analysis, in order to extract relevant information from the complex chromatogram, such as important indexes involved in the environmental studies. In particular, the project involved the identification and the characterization of homologous series of organic compounds (n-alkanes and carboxylic acids) that could be usually found in environmental samples, because they contain fundamental information to distinguish, for example, different types of emission sources, anthropic or biogenic. It has been developed a chemometric approach, which uses the AutoCoVariance Function (ACVF) computed on the digitized chromatogram, in order to quantificate the number of terms of the homologous series (nmax) and their distribution, with particular attention to the relative abundance and, consequently, the prevalance of the odd to even terms of the series (CPI). This is one of the most important parameters (environmental biomarkers) to perform a study of source apportionment. The method has been validated using simulated chromatograms and its applicability has been tested, with successful results, on real samples of known origin (e.g. gasoil or plant samples) and, finally, to particulate matter samples, obtained thanks to a collaboration with the research group of Environmental Sciences Department of the University of Milano Bicocca.

Near infrared (NIR) spectroscopy is being developed on living tissue models for noninvasively measuring in vivo glucose concentrations in individuals with diabetes. Multivariate calibration models have been built and the selectivity of each multivariate signature has been evaluated by several means. The primary objective of the research detailed in this dissertation is to practically apply noninvasive NIR glucose measurements on animal models for both short-term and long-term studies and preview future human subject evaluations. In the animal study, living tissue spectra were collected through a modified optical interface with hyper- and hypo-glycemia control. Selective measurements of glucose molecules are illustrated by the partial lease squares (PLS) algorithm, net analyte signal (NAS) vector, and hybrid linear analysis (HLA). Each model demonstrates the ability to predict prospective glucose concentrations in the short term. A restraint platform was developed for the long-term study on conscious animals. Conscious animal spectra were collected on multiple days. The anesthetized animal experiment follows on the final day. Principal component analysis (PCA) of spectra collected on different days demonstrates no significant difference between conscious animal spectra and anesthetized animal spectra. Moreover, an NAS vector analysis from conscious animal spectra has the ability predict glucose concentrations which follow the blood glucose transient during the anesthetized animal experiment. This procedure has great potential to be applied in future NIR glucose monitoring device. Before the application of this noninvasive NIR technology on people with diabetes, the impact of skin difference must be determined. In this human subject study, human skin color and baseline spectra were collected and analyzed to determine differences among individuals and within groups of people. To compare in vivo NIR spectra with different skin characteristics, PCA was performed to obtain principal component (PC) scores. Poor correlation between PC scores and skin characteristics concludes that noninvasive near-infrared technology is insensitive to different types of skin. In addition, glucose prediction was performed by a NAS analysis. The prediction results demonstrate that it is feasible to build a NAS glucose model for noninvasive NIR glucose predictions in human subjects.

“Food quality” is a very wide issue and sometimes it is also difficult to define. Food is generally a complex mixture of chemical compounds and physical properties which make up its characteristics. The determination of quality of such complicated systems is usually very difficult considering that the available analytical techniques can evaluate one constituent at a time and they cannot provide a response useful to define the general concept of “quality”. Furthermore, these techniques are often very expensive. They need sophisticated instrumentations and trained analysts. Near infrared spectroscopy (NIRS) seems to be able to solve the major part of these problems. The spectra obtained by scanning samples are like a fingerprint of the organic matter and they reveal much information about the composition and the chemical-physical properties of the scanned compounds. The real benefit of NIRS is its speed in providing responses, which is an essential feature especially for highly perishable food. NIRS has also the important advantage to be perfectly adaptable to modern food production systems where all the different steps of the supply chain are often automated and where the quality of the products has to be checked directly on the production line in order to obtain standardised products with specific characteristics. Even if NIRS is nowadays largely used by industries and laboratories, there is still space for further research in this field; through my PhD thesis I have tried to increase the applications of NIRS to solve specific problems of the food sector. The present thesis has been divided into 7 different chapters: The first one regards the fundamental aspects of NIRS. It provides a brief description of its foundation, the equipment it uses and the mathematical tools needed to extract the whole information contained in NIR spectra. Furthermore a selected number of NIRS applications have been described after conducting a bibliographic research. The chemometrics section is intended as a sort of introduction for beginners who approach this discipline for the first time. In fact it offers an overview about chemometrics without giving to many details of this large subject. The number of formulas in this thesis were kept to a minimum reporting just some calculations necessary to understand how to obtain a good calibration and how to evaluate its performances. The 6 other chapters of this thesis report the 6 different trials I carried out during my 3 years of PhD: - in the first study NIRS was used to evaluate meat quality of different rabbits’ genetic lines (using fresh instead of freeze-dried meat samples) and to discriminate rabbits according to their genetic belonging. It was demonstrated that NIRS may be a useful tool to monitor meat quality during the development of rabbit selection programs especially for the prediction of the main chemical constituents and the fatty acid profile; - the main aims of the second study were to test the potential of NIR spectroscopy to predict important composition parameters of Pecorino Siciliano cheese and to classify samples according to their aging period. Pecorino Siciliano cheese has to respect some prerequisites to be commercialized with the P.D.O. (Protected Designation of Origin) denomination: it must have a minimum amount of fat of 40% on dry matter bases and it has to be aged for at least 4 months. NIRS seems to be a strategic tool to monitor the production supply chain of traditional products so that improve and standardize their overall quality; - the objective of the 3rd and 4th studies was to compare different spectral regions, NIR (Near Infrared Reflectance) and visible-NIR (Vis-NIR), to assess their ability to discriminate between fresh and frozen-thawed fish products: sea bream fillets and swordfish cutlets. The substitution of thawed products labelled as fresh is a common commercial fraud. Consumers and honest traders must be protected from this fraudulent behaviour. Spectroscopy seems to be a really useful analytical technique able to discriminate samples which have undergone different storage treatments and it is especially useful for fish which has lost its original wholeness; - the 5th study highlights a very important topic about NIR spectroscopy: the calibration transfer amongst different instruments. A large and robust calibration for pork fat composition was available in our lab using a Foss NIRSystem 5000 as spectrophotometer. We tried to transfer this calibration to use it on two other instruments (Unity Scientific 2500x and Zeiss MMS1 sensor) so that all the information acquired during many years of lab analysis are kept. To avoid full recalibration of the new NIR instruments, a certain number of chemometrics tools are available; they permit the calibration transfer and spectra correction from instrumental and environmental differences. Several standardisation approaches have been proposed in literature and the aim of our study was to evaluate which of them provide the best performances for this purpose; - the aim of the last study was to discriminate Shiraz wines (vintage 2006) produced in 5 Australian regions (Barossa Valley, Coonawarra, McLaren Vale, Clare Valley, Western Australia) using UV-Visible (UV-Vis), Near Infrared (NIR) and Mid Infrared (MIR) spectroscopy, combined with chemometrics. Knowing the origin of food is extremely important to safeguard traditional products from illegal imitations and to guarantee “food traceability” which is nowadays an important concept worldwide. NIRS has found a strategic role in this field for its classification ability starting from the information embedded into the spectra. As many spectroscopy users would say, “NIRS and chemometrics are not a magic box where all the questions are answered and all problems find resolution” but we tried to underline the concept of versatility of this technique which can be applied to solve different problems in the whole food supply chain. On the other hand we have not hidden the limits that sometimes this secondary analytical technique can present.
La qualità degli alimenti è un concetto molto ampio e talvolta difficile da definire. Il cibo è un miscuglio di composti chimici e proprietà fisiche che insieme definiscono le caratteristiche dell’alimento stesso. Determinare la qualità di un sistema così complesso è spesso molto difficile, soprattutto considerate le tecniche analitiche disponibili, che generalmente riescono a valutare un costituente o una proprietà alla volta senza essere in grado di esprimere un giudizio generale sulla qualità del prodotto. Queste tecniche analitiche, inoltre, sono solitamente costose in quanto richiedono strumentazioni elaborate e personale addetto qualificato. La spettroscopia nel vicino infrarosso (NIRS) sembra essere in grado di risolvere buona parte di queste problematiche. Gli spettri ottenuti attraverso lo scansionamento dei campioni possono essere considerati una sorta di “impronta digitale” della sostanza organica e possono contenere molte informazioni riguardanti la composizione e le proprietà chimico-fisiche degli alimenti. Il vero vantaggio della tecnologia NIR resta comunque la sua velocità nel fornire i risultati. Questa caratteristica è estremamente importante nel settore alimentare, dove i prodotti sono altamente deperibili e devono essere commercializzati nel più breve tempo possibile. La spettroscopia NIR si adatta perfettamente alle moderne esigenze delle industrie alimentari dove le catene di produzione sono principalmente automatizzate e la qualità degli alimenti deve essere monitorata costantemente in modo tale da standardizzare la qualità del prodotto finito. Nonostante la spettroscopia NIR venga utilizzata da laboratori ed industrie alimentari da parecchi anni, c’è ancora spazio per ulteriore ricerca in questo campo. Attraverso la mia tesi di dottorato ho cercato di ampliare le possibili applicazioni della tecnologia NIR per risolvere delle problematiche concrete del settore “produzione e commercializzazione” degli alimenti. Il lavoro è stato suddiviso in 7 capitoli. Il primo riguarda alcuni aspetti generali della spettroscopia NIR e fornisce una breve descrizione dei fondamenti, delle attrezzature e degli strumenti necessari per l‘interpretazione degli spettri. La parte riguardante la chemiometria vuole essere una sorta di introduzione per coloro che si avvicinano per la prima volta a questa disciplina molto ampia. Il numero di formule inserito in questa tesi è stato mantenuto il più basso possibile, riportando solo quelle necessarie per ottenere delle buone calibrazioni e per poter poi testare le loro performance. Gli altri sei capitoli della tesi trattano i diversi contributi sperimentali che ho condotto durante i miei tre anni di dottorato di ricerca: - nel primo studio la spettroscopia NIR è stata utilizzata per valutare la qualità della carne di coniglio basando l’analisi sul prodotto fresco. È stata inoltre testata la capacità discriminante di questa tecnologia per classificare i campioni sulla base della loro linea genetica di appartenenza. È stato dimostrato che la tecnologia NIR potrebbe diventare una tecnica analitica di supporto nei piani di selezione genetica dei conigli per monitorare la qualità della loro carne soprattutto per quel che riguarda la predizione dei principali costituenti chimici e del profilo acidico di questo prodotto; - gli obiettivi del secondo contributo sperimentale sono stati la predizione della composizione centesimale del formaggio Pecorino Siciliano DOP e la classificazione di questo prodotto sulla base del suo grado di stagionatura. Per essere commercializzato con la denominazione DOP, questo formaggio deve rispettare alcuni prerequisiti minimi quali un contenuto di grasso pari al 40% sulla sostanza secca e un livello di stagionatura minima di 4 mesi. NIRS sembra una tecnica analitica molto utile per poter standardizzare la qualità e le tecniche di produzione dei prodotti tradizionali italiani che per essere commercializzati su mercati nazionali ed internazionali devono presentare caratteristiche ben definite; - l’obiettivo del terzo e quarto contributo sperimentale è stato il confronto di diverse regioni spettrali quali il NIR (vicino infrarosso) e il visibile-vicino infrarosso (Vis-NIR) per la discriminazione di pesce fresco da quello decongelato. I due contributi avevano come oggetto di studi due prodotti ittici di qualità quali i filetti di orata e i tranci di pesce spada. La vendita di prodotto decongelato spacciato per fresco è una comune frode commerciale da cui i consumatori e i commercianti devono essere protetti. La spettroscopia si è dimostrata una tecnica analitica utile per discriminare campioni che hanno subito trattamenti di conservazione diversi, in particolare per quei prodotti ittici che hanno perso la loro integrità anatomica (quali filetti e tranci); - il quinto studio ha trattato un argomento particolarmente importante nell’ambito della spettroscopia NIR: il trasferimento di calibrazioni tra diversi strumenti. Il nostro laboratorio aveva a disposizione una calibrazione ampia e robusta costruita sul grasso suino per la predizione del numero di iodio e del profilo acidico. Le equazioni erano state ottenute usando uno spettrofotometro FOSS NIRSystem 5000. Il nostro obiettivo era quello di trasferire questa calibrazione su altri due strumenti (Unity Scientific 2500x e un sensore Zeiss MMS1 portatile) per poter quindi mantenere e sfruttare l’informazione acquisita durante molti anni di analisi chimiche di laboratorio su questa matrice. Per evitare la ricalibrazione degli strumenti secondari, la chemiometria mette a disposizione diverse tecniche definite “modelli di standardizzazione” che nel nostro studio sono state poste a confronto; - l’obiettivo dell’ultimo contributo introdotto in questa tesi è stata la discriminazione di vino Shiraz prodotto in 5 diverse regioni australiane particolarmente vocate per il settore vitivinicolo (Barossa Valley, McLaren Vale, Clare Valley, Coonawarra, Western Australia). Diverse regioni spettrali sono state confrontate a tale scopo (Visibile, vicino infrarosso e medio infrarosso), sfruttando inoltre diverse tecniche di classificazione che la chemiometria mette a disposizione. Conoscere l’origine degli alimenti è di particolare importanza soprattutto per salvaguardare i prodotti tradizionali. Questi, infatti, rischiano di essere sostituiti con prodotti simili ma di qualità inferiore. La spettroscopia NIR, impiegando le informazioni contenute nello spettro di un alimento, si è rivelata in grado di discriminare i prodotti sulla base della loro origine. Sicuramente la spettroscopia NIR non è una “scatola magica” dove tutte le domande trovano risposta e i problemi vengono risolti, ma possiede una versatilità tale da consentirne l’impiego in molteplici realtà industriali e commerciali. Attraverso la mia tesi ho cercato di mettere in luce le potenzialità di questa tecnologia senza però nascondere i limiti che ho riscontrato nel corso del suo utilizzo.

Proteomics is a rapidly growing discipline dealing with structure, molecular interactions, conformational dynamics, modifications and the functions of proteins. Mass spectrometry (MS) and (nuclear magnetic resonance) (NMR) have been used comprehensively to study protein interactions. Acyl carrier protein (ACP) interacts with more than 30 partner proteins during either fatty acid or polyketide biosynthesis. In order to be fully activated ACP gains a 4'-phosphopantetheinyl (4'-PP) group from coenzyme A using acyl carrier protein synthase (AcpS) via posttranslational modification. Protein-protein interactions of the ACP from the actinorhodin (act) polyketide synthase (PKS) complex and AcpS were investigated using oxidative footprinting with hydroxyl radicals generated from the Fenton reaction. Chemical modification of acidic residues was also used to investigate the interaction between these proteins using l-ethyl-3-(3- dimethylaminopropyl) carbodiimide (EDC) this acted as a zero length cross linker to induce modification with Me-Glycine. MS was used to identify the modified residues and peptides and the extent of modification. Several residues were found to be protected in the complex between the two proteins and these may participate in the interaction interface between ACP and ACPS. Isotopically labelled ACP was expressed and purified and multidimensional NMR experiments were recorded to investigate this interaction interface identified using oxidative footprinting techniques. Chemical shift perturbations for ACP residues were calculated, and these revealed that many residues were affected by oxidation of ACP. Oxidation of methionine to methionine sulfoxide was confirmed. Metabolomics is a discipline which deals with metabolites in a biological system. It provides a wealth information for disease diagnosis, drug discovery, toxicology and genetic modification. Attempts have been made in this thesis to utilize metabolomics in biometrics. Mice were used as a model to attempt to determine individuals' age by their scent. In this part of the project chemometric methods were used to discriminate mice using a gas chromatography-mass spectrometry dataset of volatile organic compounds obtained from their urine. Principal component regression (PCR), partial least squares regression (PLSR) and support vector regression (SVR) were used to determine mouse age. Mice could be discriminated by their age using SVR without overfitting the data.

Questo lavoro di tesi risponde al bisogno dell’industria lattiero-casearia di incrementare la produttività e allo stesso tempo soddisfare la richiesta dei consumatori di prodotti di elevata qualità. A tal fine, si possono proporre alle aziende lattiero-casearie nuovi metodi per migliorare la comprensione e il monitoraggio dei processi di produzione. La Process Analytical Technology (PAT) rappresenta uno strumento ideale per raggiungere questo scopo, grazie a sensori in grado di eseguire analisi rapide, green, non distruttive ed in tempo reale. Le tecniche più utilizzate in questo campo sono la spettroscopia del vicino e del medio infrarosso (NIR e MIR, rispettivamente), che forniscono informazioni chimico-fisiche sul prodotto grazie a sonde installate direttamente in punti critici del processo. Tuttavia, queste tecniche hanno lo svantaggio di fornire risultati (spettri) difficilmente interpretabili senza l’aiuto di un adeguato metodo statistico. In questo contesto, gli algoritmi Chemiometrici permettono l’estrazione di informazioni rilevanti dai dati spettroscopici, permettendo la comprensione del sistema studiato. La prima parte del presente lavoro è focalizzata sul monitoraggio del processo di coagulazione, uno dei momenti più critici della caseificazione. A tal fine, è stata utilizzata una sonda FT-NIR per acquisire spettri durante il processo di coagulazione. variando alcuni fattori tecnologici cruciali, come temperatura, contenuto di grasso e pH, secondo un disegno Box-Behnken. Attraverso l’algoritmo Multivariate Curve Resolution - Alternating Least Squares (MCR-ALS) è stato possibile ottenere sia una efficiente descrizione delle tre differenti fasi del processo di coagulazione, sia lo sviluppo di carte di controllo multivariate (Multivariate Statistical Process Control charts, MSPC), capaci di individuare possibili non-conformità fin dai primi momenti del processo. Inoltre, il metodo ANOVA-Simultaneous Component Analysis (ASCA) è stato applicato ai dati spettrali al fine di ottenere una migliore comprensione della coagulazione, evidenziando in che modo ogni fattore sperimentale influenzi il processo. Nella seconda parte del lavoro, la spettroscopia FT-NIR è stata studiata come possibile strumento per sostituire le tecniche standard, come il Formagraph, per valutare l’attitudine alla coagulazione. Le prove di coagulazione sono state effettuate usando differenti campioni di latte in polvere. L’utilizzo dell’algoritmo MCR-ALS ha permesso la valutazione della miglior polvere in termini di attitudine alla coagulazione e, inoltre, ha evidenziato la non significatività degli effetti della concentrazione di CaCl2 e del trattamento termico del latte ricostituito sul tempo di coagulazione. Infine, le prove sperimentali eseguite con miscele di latte scremato e percentuali più elevate di latte ricostituito hanno mostrato una coagulazione più lenta. L’ultima parte del lavoro ha riguardato l’utilizzo della spettroscopia MIR per monitorare la produzione di galattooligosaccaridi (GOS) dal siero di formaggio, allo scopo di valorizzare questo prodotto e di ottimizzare il processo. La regressione Partial Least Square (PLS) è stata utilizzata con l’obiettivo di predire le componenti specifiche derivanti dalle differenti reazioni enzimatiche studiate. In conclusione, l’applicazione dei metodi proposti permetterà un efficiente controllo del processo garantendo un modesto impatto ambientale e soddisfacendo allo stesso tempo requisiti di legge e esigenze dei consumatori. Infine, l’affidabilità degli approcci PAT può essere rafforzata da future applicazioni industriali.
This thesis work wants to answer the need of dairy industry to increase productivity while satisfying the consumers request for higher quality products. In order to do that, dairy companies need innovative methods to improve the understanding and the monitoring of production processes. Process Analytical Technology (PAT) approaches are the perfect tool for this purpose, as they use green, fast, non-invasive and non-destructive sensors that allow to perform measurements in real time. The most used techniques in this field are Near- and Mid-Infrared (NIR and MIR, respectively) spectroscopy, whose probes can be directly installed in critical points of the process providing both physical and chemical information of the product. However, these techniques have the drawback of providing results (spectra) difficult to be interpreted without proper statistical tools. In this context, Chemometric methods and algorithms allow the extraction of relevant information from spectroscopic data, providing a better understanding of the studied system. The first part of the present work focused on the monitoring of the coagulation process, one of the most critical moments of cheesemaking. To this aim, an FT-NIR spectroscopy system was used, acquiring spectra along the rennet coagulation process. According to a Box-Behnken experimental design, several coagulation trials were carried out, changing crucial technological factors, such as temperature, fat content and pH. Through Multivariate Curve Resolution optimized by Alternating Least Squares (MCR-ALS) algorithm it was possible to both have a reliable description of the three different coagulation phases and, most importantly, to build Multivariate Statistical Process Control (MSPC) charts, able to detect failures from the first moment of the process. Moreover, ANOVA-Simultaneous Component Analysis (ASCA) method was applied on spectral data to obtain a better understanding of the process, highlighting in which way each physicochemical parameter affects the process. In the second part of the work, FT-NIR spectroscopy was tested as a possible tool to replace the golden standards of coagulation ability, i.e. Formagraph. Coagulation trials were carried out using different milk powder samples. The use of MCR-ALS algorithm permitted the assessment of the best powder in terms of coagulation attitude and, in addition, it highlighted the non-significant effect on coagulation occurrence of CaCl2 concentration and of heat treatment on reconstituted milk. Finally, experimental trials carried out with mixtures of skimmed milk and reconstituted milk showed a slower coagulation time when a higher reconstituted milk percentage was used. The last part of the work regarded the use of MIR spectroscopy to monitor Galactooligosaccharides (GOS) production from cheese whey, in order to avoid the waste of this compound and to optimize the studied process. To do so, Partial Least Square (PLS) regression was used to predict the specific compounds resultant from the different enzymatic reaction studied. In conclusion, the application of the proposed methods will implicate, with a modest environmental impact, an efficient control of the process, satisfying at the same time law requirements and consumers’ needs. Furthermore, reliability of PAT approaches could be strengthened by future industrial applications.

Interest in food quality and production has increased in recent decades, mainly due to changes in consumer habits and behaviour, and the development and increase in the industrialisation of food chains. The growing demand for quality and safety in food production obviously calls for high standards for quality and process control, which in turn requires appropriate analytical tools for the analysis of food. In particular, many unit operations in industrial food processes are related to microbial fermentation, namely milk coagulation in dairy, dough in bakery, as well as must fermentation in wine and beer productions. Fermentation is one of the earliest methods adopted to obtain value-added food products with an extended shelf life. Humans applied fermentation to make products such as wine, mead, cheese and beer long before the biochemical process behind was understood. Even now the biochemistry of fermentations commonly applied in food processes has many aspects which have not been fully investigated yet. Briefly, fermentation is any metabolic process in which an organism converts a carbohydrate, such as starch or sugar, into an alcohol and/or organic acids entailing modifications in the final product. The transition to industrial productions entailed a standardisation of the fermentation processes and the obtained products. Currently, the main objective is to develop instruments able to be implemented in the process in order to closely monitor the products of interest and to detect in real time the smallest changes bringing to a more effective process control and management. In this contest, spectroscopy revealed to be an interesting analytical method to monitor food fermentations processes. Spectroscopy is a secondary analytical method which consists in recording the absorption changes due to the interaction of electromagnetic radiation with the matter. The basic principle is that every chemical compound absorbs, transmits or reflects light (electromagnetic radiation) over a certain range of wavelengths. The information recorded can, thus, be used to measure the amount of a known chemical substance if correlated to a reference analysis. Spectroscopy reveals to be one of the most useful methods for quantitative analysis in various fields such as chemistry, physics, biochemistry, material and chemical engineering and clinical applications. Indeed, any application that deals with chemical substances or materials can use this technique. Moreover, the improved instrumentation for performing in-line and on-line analyses at industrial level has rose in the last decades giving the opportunity to obtained real-time information about the progression of any process and allowed its implementation as strategy to monitor complex systems as food production. The food monitoring with spectroscopic devices has become possible thanks to Chemometrics (i.e. multivariate data analysis). Chemometrics has widely demonstrated to be the perfect partner to spectroscopy to deal with the complex chemical/physical systems that food matrix conforms. Chemometrics is able to extract relevant information from redundant and noisy spectra. In the last years the combination of spectroscopic analysis and Chemometrics was applied crosswise in food processes for qualitative and quantitative modelling in industrial applications. In particular, for the determination of compositional parameters affecting quality and safety of fermented food products such as wine, beer, yoghurt, vinegar and bakery products. Nevertheless, concerning complex biotransformations spectroscopy and Chemometrics are emerging techniques in food fermentation monitoring. The purpose of this PhD Thesis is the demonstration of the feasibility in the combination of spectroscopy and Chemometrics as an innovative working procedure for real time monitoring of food fermentation processes. The thesis consists of five main chapters Chapter 1 Chapters 2 and 3 present an introduction to the main fermentations and their control from an historical prospective, the employed analytical techniques (Near infrared and Mid Infrared spectroscopy) and to Chemometrics, respectively. Chapter 4 presents the experiments carried out on various fermentation food processes. In this section seven studies represent examples of applications of different spectroscopic methods in strong combination with Chemometrics to food fermentation processes as yogurt fermentation (Paper I, II and Paper III), wine malolactic transformation (Paper IV and V) and beer (Paper VI and VII). In addition to the mentioned contributions a brief state of the art and some preliminary results are reported regarding sourdough leaving process monitoring. The two basic Chemometrics tools, principal component analysis (PCA) and partial least squares (PLS) regression were mainly applied to the spectroscopic data collected from the fermentation processes in order to evaluate the results and focus on the relevant information and to correlate the spectral features with different relevant physical and/or chemical parameters such as the concentration of the main chemical species involved in the biotransformation. In particular, the principal components (PCs) scores obtained by monitoring wine and yoghurt fermentations were modelled as function of time to find out kinetic parameters, as maximum acceleration and deceleration of the transformation, important for the process control (PAPER I and V). The spectroscopic data obtained during yoghurt and beer fermentation monitoring were also investigated with multivariate curve resolution- alternating least squares (MCR-ALS), proving to be able to resolve multi-component mixtures into a simpler model (PAPER II and VII). The main conclusive remarks on the presented studies are given in Chapter 5 (CONCLUSIONS), including a discussion of challenges and future perspectives for further application of spectral monitoring and chemometrics in fermented food processes.

Human hair profiling is an integral part of a forensic investigation but it is one of the most technically difficult subjects in forensic science. This thesis describes the research and development of a novel approach for the rapid identification of unknown human and other related keratin fibres found at a crime scene. The work presented here is developed systematically and considers sample collection, sample preparation, analysis and interpretation of spectral data for the profiling of hair fibres encountered in criminal cases. Spectral comparison of fibres was facilitated with the use of chemometrics methods such as PCA, SIMCA and Fuzzy Clustering, and the less common approach of multi-criteria decision making methodology (MCDM). The aim of the thesis was to investigate the potential of some vibrational spectroscopy techniques for matching and discrimination of single keratin hair fibres in the context of forensic evidence. The first objective (chapter 3) of the thesis was to evaluate the use of Raman and FT-IR micro-spectroscopy techniques for the forensic sampling of hair fibres and to propose the preferred technique for future forensic hair comparisons. The selection of the preferred technique was based on criteria such as spectral quality, ease of use, rapid analysis and universal application to different hair samples. FT-IR micro-spectroscopy was found to be the most appropriate technique for hair analysis because it enabled the rapid collection of spectra from a wide variety of hair fibres. Raman micro-spectroscopy, on the other hand, was hindered with fluorescence problems and did not allow the collection of spectra from pigmented fibres. This objective has therefore shown that FT-IR micro-spectroscopy is the preferable spectroscopic technique for forensic analysis of hair fibres, whilst Raman spectroscopy is the least preferred. The second objective (chapter 3) was to investigate, through a series of experiments, the effect of chemical treatment on the micro-environment of human hair fibres. The effect of bleaching agents on the hair fibres was studied with some detail at different treatment times and the results indicate a significant change in the chemical environment of the secondary structure of the hair fibre along with changes in the C-C backbone structure. One of the most important outcomes of this research was the behaviour of the fÑ-helix during chemical treatment. The hydrogen bonding in the fÑ-helix provides for the stable structure of the fibre and therefore any disruption to the fÑ-helix will inevitably damage the molecular structure of the fibre. The results highlighted the behaviour of the fÑ-helix, which undergoes a significant decrease in content during oxidation, and is partly converted to a random-coil structure, whilst the fÒ-sheet component of the secondary structure remains unaffected. The reported investigations show that the combination of FT-IR and Raman micro-spectroscopy can provide an insight and understanding into the complex chemical properties and reactions within a treated hair fibre. Importantly, this work demonstrates that with the aid of chemometrics, it is possible to investigate simultaneously FT-IR and Raman micro-spectroscopic information from oxidised hair fibres collected from one subject and treated at different times. The discrimination and matching of hair fibres on the basis of treatment has potential forensic applications. The third objective (chapter 4) attempted to expand the forensic application of FT-IR micro-spectroscopy to other keratin fibres. Animal fibres are commonly encountered in crime scenes and it thus becomes important to establish the origin of those fibres. The aim of this work was to establish the forensic applications of FT-IR micro-spectroscopy to animal fibres and to investigate any fundamental molecular differences between these fibres. The results established a discrimination between fibres consisting predominantly of fÑ-helix and those containing mainly a fÒ-sheet structure. More importantly, it was demonstrated through curve-fitting and chemometrics, that each keratin fibre contains a characteristic secondary structure arrangement. The work presented here is the first detailed FT-IR micro-spectroscopic study, utilising chemometrics as well as MCDM methods, for a wide range of keratin fibres, which are commonly, found as forensic evidence. Furthermore, it was demonstrated with the aid of the rank ordering MCDM methods PROMETHEE and GAIA, that it is possible to rank and discriminate keratin fibres according to their molecular characteristics obtained from direct measurements together with information sourced from the literature. The final objective (chapter 5) of the thesis was to propose an alternative method for the discrimination and matching of single scalp human hair fibres through the use of FT-IR micro-spectroscopy and chemometrics. The work successfully demonstrated, through a number of case scenarios, the application of the technique for the identification of variables such as gender and race for an unknown single hair fibre. In addition, it was also illustrated that known hair fibres (from the suspect or victim) can be readily matched to the unknown hair fibres found at the crime scene. This is the first time that a substantial, systematic FT-IR study of forensic hair identification has been presented. The research has shown that it is possible to model and correlate individual¡¦s characteristics with hair properties at molecular level with the use of chemometrics methods. A number of different, important forensic variables of immediate use to police in a crime scene investigation such as gender, race, treatment, black and white hair fibres were investigated. Blind samples were successfully applied both to validate available experimental data and extend the current database of experimental determinations. Protocols were posed for the application of this methodology in the future. The proposed FT-IR methodology presented in this thesis has provided an alternative approach to the characterisation of single scalp human hair fibres. The technique enables the rapid collection of spectra, followed by the objective analytical capabilities of chemometrics to successfully discriminate animal fibres, human hair fibres from different sources, treated from untreated hair fibres, as well as black and white hair fibres, on the basis of their molecular structure. The results can be readily produced and explained in the courts of law. Although the proposed relatively fast FT-IR technique is not aimed at displacing the two slower existing methods of hair analysis, namely comparative optical microscopy and DNA analysis, it has given a new dimension to the characterisation of hair fibres at a molecular level, providing a powerful tool for forensic investigations.

Human hair profiling is an integral part of a forensic investigation but it is one of the most technically difficult subjects in forensic science. This thesis describes the research and development of a novel approach for the rapid identification of unknown human and other related keratin fibres found at a crime scene. The work presented here is developed systematically and considers sample collection, sample preparation, analysis and interpretation of spectral data for the profiling of hair fibres encountered in criminal cases. Spectral comparison of fibres was facilitated with the use of chemometrics methods such as PCA, SIMCA and Fuzzy Clustering, and the less common approach of multi-criteria decision making methodology (MCDM). The aim of the thesis was to investigate the potential of some vibrational spectroscopy techniques for matching and discrimination of single keratin hair fibres in the context of forensic evidence. The first objective (chapter 3) of the thesis was to evaluate the use of Raman and FT-IR micro-spectroscopy techniques for the forensic sampling of hair fibres and to propose the preferred technique for future forensic hair comparisons. The selection of the preferred technique was based on criteria such as spectral quality, ease of use, rapid analysis and universal application to different hair samples. FT-IR micro-spectroscopy was found to be the most appropriate technique for hair analysis because it enabled the rapid collection of spectra from a wide variety of hair fibres. Raman micro-spectroscopy, on the other hand, was hindered with fluorescence problems and did not allow the collection of spectra from pigmented fibres. This objective has therefore shown that FT-IR micro-spectroscopy is the preferable spectroscopic technique for forensic analysis of hair fibres, whilst Raman spectroscopy is the least preferred. The second objective (chapter 3) was to investigate, through a series of experiments, the effect of chemical treatment on the micro-environment of human hair fibres. The effect of bleaching agents on the hair fibres was studied with some detail at different treatment times and the results indicate a significant change in the chemical environment of the secondary structure of the hair fibre along with changes in the C-C backbone structure. One of the most important outcomes of this research was the behaviour of the fÑ-helix during chemical treatment. The hydrogen bonding in the fÑ-helix provides for the stable structure of the fibre and therefore any disruption to the fÑ-helix will inevitably damage the molecular structure of the fibre. The results highlighted the behaviour of the fÑ-helix, which undergoes a significant decrease in content during oxidation, and is partly converted to a random-coil structure, whilst the fÒ-sheet component of the secondary structure remains unaffected. The reported investigations show that the combination of FT-IR and Raman micro-spectroscopy can provide an insight and understanding into the complex chemical properties and reactions within a treated hair fibre. Importantly, this work demonstrates that with the aid of chemometrics, it is possible to investigate simultaneously FT-IR and Raman micro-spectroscopic information from oxidised hair fibres collected from one subject and treated at different times. The discrimination and matching of hair fibres on the basis of treatment has potential forensic applications. The third objective (chapter 4) attempted to expand the forensic application of FT-IR micro-spectroscopy to other keratin fibres. Animal fibres are commonly encountered in crime scenes and it thus becomes important to establish the origin of those fibres. The aim of this work was to establish the forensic applications of FT-IR micro-spectroscopy to animal fibres and to investigate any fundamental molecular differences between these fibres. The results established a discrimination between fibres consisting predominantly of fÑ-helix and those containing mainly a fÒ-sheet structure. More importantly, it was demonstrated through curve-fitting and chemometrics, that each keratin fibre contains a characteristic secondary structure arrangement. The work presented here is the first detailed FT-IR micro-spectroscopic study, utilising chemometrics as well as MCDM methods, for a wide range of keratin fibres, which are commonly, found as forensic evidence. Furthermore, it was demonstrated with the aid of the rank ordering MCDM methods PROMETHEE and GAIA, that it is possible to rank and discriminate keratin fibres according to their molecular characteristics obtained from direct measurements together with information sourced from the literature. The final objective (chapter 5) of the thesis was to propose an alternative method for the discrimination and matching of single scalp human hair fibres through the use of FT-IR micro-spectroscopy and chemometrics. The work successfully demonstrated, through a number of case scenarios, the application of the technique for the identification of variables such as gender and race for an unknown single hair fibre. In addition, it was also illustrated that known hair fibres (from the suspect or victim) can be readily matched to the unknown hair fibres found at the crime scene. This is the first time that a substantial, systematic FT-IR study of forensic hair identification has been presented. The research has shown that it is possible to model and correlate individual¡¦s characteristics with hair properties at molecular level with the use of chemometrics methods. A number of different, important forensic variables of immediate use to police in a crime scene investigation such as gender, race, treatment, black and white hair fibres were investigated. Blind samples were successfully applied both to validate available experimental data and extend the current database of experimental determinations. Protocols were posed for the application of this methodology in the future. The proposed FT-IR methodology presented in this thesis has provided an alternative approach to the characterisation of single scalp human hair fibres. The technique enables the rapid collection of spectra, followed by the objective analytical capabilities of chemometrics to successfully discriminate animal fibres, human hair fibres from different sources, treated from untreated hair fibres, as well as black and white hair fibres, on the basis of their molecular structure. The results can be readily produced and explained in the courts of law. Although the proposed relatively fast FT-IR technique is not aimed at displacing the two slower existing methods of hair analysis, namely comparative optical microscopy and DNA analysis, it has given a new dimension to the characterisation of hair fibres at a molecular level, providing a powerful tool for forensic investigations.

Human health is a complex and wide-ranging subject far beyond nutrition and physical exercise. Still, these factors have a huge impact on global health by their ability to prevent diseases and thus promote health. Thus, to identify health risks and benefits, it is necessary to reveal the underlying mechanisms of nutrition and exercise, which in many cases follows a complex chain of events. As a consequence, current health research is generating massive amounts of data from anthropometric parameters, genes, proteins, small molecules (metabolites) et cetera, with the intent to understand these mechanisms. For the study of health responses, especially related to physical exercise and nutrition, alterations in small molecules (metabolites) are in most cases immediate and located close to the phenotypic level and could therefore provide early signs of metabolic imbalances. Since there are roughly as many different responses to exercise and nutrients as there are humans, this quest is highly multifaceted and will benefit from an interpretation of treatment effects on a general as well as on an individual level. This thesis involves the application of chemometric methods to the study of global metabolic reactions, i.e. metabolomics, in a strategy coined predictive metabolomics. Via the application of predictive metabolomics an extensive hypothesis-free biological interpretation has been carried out of metabolite patterns in blood, acquired using gas chromatography-mass spectrometry (GC-MS), related to physical exercise, nutrition and diet, all in the context of human health. In addition, the chemometrics methodology have computational benefits concerning the extraction of relevant information from information-rich data as well as for interpreting general treatment effects and individual responses, as exemplified throughout this work. Health concerns all lifestages, thus this thesis presents a strategic framework in combination with comprehensive interpretations of metabolite patterns throughout life. This includes a broad range of human studies revealing metabolic patterns related to the impact of physical exercise, macronutrient modulation and different fitness status in young healthy males, short and long term dietary treatments in overweight post menopausal women as well as metabolic responses related to probiotics treatment and early development in infants. As a result, the studies included in the thesis have revealed metabolic patterns potentially indicative of an anti-catabolic response to macronutrients in the early recovery phase following exercise. Moreover, moderate differences in the metabolome associated with cardiorespiratory fitness level were detected, which could be linked to variation in the inflammatory and antioxidaive defense system. This work also highlighted mechanistic information that could be connected to dietary related weight loss in overweight and obese postmenopausal women in relation to short as well as long term dietary effects based on different macronutrient compositions. Finally, alterations were observed in metabolic profiles in relation to probiotics treatment in the second half of infancy, suggesting possible health benefits of probiotics supplementation at an early age.
Embargo until 2012-06-01

Ce travail a consisté en la recherche et l’exploitation de nouveaux motifs structuraux sensibles à des médiateurs chimiques. Pour cela, une approche chimiométrique a été développée dans le but d’obtenir des profils de réactivité pouvant être classés dans un tableau avec un code de couleur afin de pouvoir mettre en évidence visuellement les motifs présentant un potentiel intéressant. Les motifs d’intérêt ont été intégrés dans des sondes pro-fluorescentes qui ont ensuite été testées en milieux biologiques afin d’observer leur activité. Cette méthodologie a permis de révéler un nouveau type de quencher biologiquement et chimiquement désactivable. Ainsi, même en l’absence du médiateur étudié, ce quencher incorporé dans une sonde de type FRET sera réactivé par ajout d’un agent chimique exogène afin de révéler les sondes non-activées dans la cellule
The main topic of this work was the research and the use of new structural patterns sensitive to chemical mediators. A chimiometric approach was developped to obtain reactivity profiles which will be filed in a table with a color code in order to visually highlight the patterns having interessant potential. Then, the patterns of interest were integrated in FRET-based probes which were tested in cell experiments. This profiling led to a new type of biologically and chemically deactivatable quencher. Thus, even in the absence of the studied mediator, this quencher incorporated in a FRET probe will be activated by adding an exogenous chemical agent to reveal inactivated probes in the cell

Due to recent progress, higher-order chemical instrumentation provides large amounts of data which need automated processing in order to extract relevant information. In most cases, the raw signals or spectra are too complex for manual analysis. The ability to detect, identify and quantitate chemical substances in gas phase in field operations is required in a huge number of applications. Among them, I would like to highlight the need for chemical sensing on diverse humanitarian, safety and security applications. In these cases, it becomes extremely important to continuously monitor the environments where chemicals are spread in order to be ready to act when abnormal events are discovered. In most critical scenarios the sample can not just be taken to the laboratory and analyzed, since an immediate answer is needed. In some other scenarios, the exploration of the area must be performed because the localization of the gas source or material of interest is unknown. This exploration can be performed using multiple mobile sensors in order to localize the chemical source or material. Different sensing technologies have been successfully used to detect and identify different chemical substances (gases or volatile compounds). These compounds could be either toxic or hazardous, or they can be signatures of the materials to be detected, for instance, explosives or drugs. Among these technologies, mobility based analyzers provide fast responses with high sensitivity. However, IMS instruments are not exempt of problems. Typically, they provide moderate selectivity, appearing overlapped peaks in the spectra. Moreover, the presence of humidity makes peaks wider, thus worsening the resolving power and the resolution. Furthermore, the response of IMS is non-linear as substance concentration increases and more than one peak can appear in the spectra due to the same compound. In the present thesis, these problems are addressed and applications using an Ion Mobility Spectrometer (IMS) and a Differential Mobility Analyzer (DMA) are shown. It is demonstrated that multivariate data analysis tools are more effective when dealing with these technologies. For the first time, multivariate data analysis tools have been applied to a novel DMA. It is shown that DMA could be established as a good instrumentation for the detection of explosives and the detection and quantitation of VOCs. Furthermore, Multivariate curve resolution Alternating Least Squares (MCR-ALS) is shown to be suitable to analyze IMS spectra qualitatively when interfering chemicals appear in the spectra and even when their behaviour is non-linear. Partial Least Squares (PLS) methods are demonstrated to work properly for the quantitative analysis of these signals; from this analysis the chemical concentrations of the target substances are obtained. It is also demonstrated in this thesis that the quantitative measurements from these sensors can be integrated in a gas source localization algorithm in order to improve the localization of the source in those scenarios where it is required. It is shown that the new proposal works significantly better in cases where the source strength is weak. This is illustrated presenting results from simulations generated under realistic conditions. Moreover, real-world data were obtained using a mobile robot mounting a photo ionization detector (PID). Experiments were carried out under forced ventilation and turbulences in indoors and outdoors environments. The results obtained validate the simulation results and confirm that the new localization algorithm can effectively operate in real environments.
Debido a los progresos recientes, la instrumentación química genera mayores volúmenes de datos los cuales requieren de un procesado automático con la finalidad de extraer la información relevante, ya que un análisis manual no suele ser viable debido a la elevada complejidad de los datos. La habilidad de detectar, identificar y cuantificar sustancias químicas en fase gas en operaciones de campo es requerida en un gran número de aplicaciones. Entre ellas, aplicaciones humanitarias y de seguridad. En estos casos, la monitorización continua de los entornos es extremadamente importante, ya que se debe estar alerta de eventos anormales. En los escenarios más críticos, debe realizarse una exploración del área porque la posición de la fuente de gas de interés es desconocida. Esta exploración puede realizarse usando múltiples robots. Diferentes tecnologías de sensores se han aplicado con éxito a la detección e identificación de diferentes sustancias químicas (gases o compuestos volátiles). Estos compuestos pueden ser tóxicos, peligrosos, o precursores de explosivos o drogas. De entre estas tecnologías, los analizadores basados en movilidad iónica (IMS) proporcionan rápidas respuestas con gran sensibilidad. Sin embargo, estos instrumentos no están exentos de problemas. Típicamente, proporcionan una moderada selectividad, apareciendo picos solapados en los espectros. Además, la presencia de humedad provoca que los picos se ensanchen, así empeorando la resolución. Además, la respuesta de IMS es no lineal al incrementar la concentración y es posible que más de un pico debido al mismo compuesto aparezca en el espectro. En la presente tesis se trata con estos problemas y se demuestra que las herramientas de análisis de datos multivariantes son más efectivas que las herramientas típicas univariantes al tratar con tecnologías de movilidad iónica (IMS y DMA), especialmente para el análisis cualitativo y cuantitativo de sus espectros. Además, se demuestra que las medidas cuantitativas pueden integrarse de manera efectiva en un algoritmo de localización de fuentes químicas. Los resultados obtenidos (simulaciones realistas y datos reales) muestran que el algoritmo desarrollado durante la tesis puede funcionar especialmente bien en situaciones en las que la potencia de emisión de la fuente a detectar sea débil.

L'huile d'olive est une composante vitale du régime méditerranéen en raison de sa valeur nutritionnelle et économique bien connues. Plusieurs facteurs environnementaux, agricoles et technologiques jouent un rôle important dans la définition de la qualité de l'huile d'olive. Au Liban, des études préliminaires ont montré que certains critères de qualité dépassent les valeurs seuils des normes éditées par le Conseil Oléicole International (COI) pour les huiles d’olive extra-vierges, dont les causes n'ont pas été identifiées. En conséquence, quatre-vingt-seize échantillons d'huile d'olive ont été récoltés sur deux saisons, traités en utilisant différentes méthodes d'extraction et collectés sur huit sites (Akkar, Chouf, Hasbaya, Koura, Tyr, Nabatiyeh, Zgharta et Hermel) à fort potentiel pour l’obtention d’une indication géographique protégée (IGP). Dans cette optique, les huiles extraites et/ou collectées, ont été soumises à une analyse chimique conventionnelle comme suggéré par le COI et à une analyse rapide en utilisant la spectroscopie de fluorescence 3D en mode frontal (3D-FFFS) et la chromatographie en phase gazeuse ultra-rapide (Ultra-Fast GC).Une corrélation entre le profil en acides gras et les conditions pédoclimatiques des principales régions oléicoles du Liban a été constatée. L'altitude, la température et l'humidité relative sont les principaux facteurs d’influence du profil d'acides gras. Les régions libanaises à haute altitude, à température moyenne basse et à faible humidité relative ont une teneur élevée en acide oléique. Les zones à basse altitude, à température moyenne plus élevée et avec une humidité relative plus élevée ont un profil en acides gras caractérisé par les acides linoléique, linolénique, palmitoléique et palmitique. Les facteurs agricoles, en particulier la date de récolte, affectent également les constituants majeurs et mineurs de l'huile d'olive. En effet, l'acidité et les polyphénols totaux étaient fortement influencés par celle-ci. De plus, une modification du profil d'acides gras caractérisée par une teneur en linoléique plus élevée, une teneur en oléique plus faible, une augmentation du ∆7-stigmasténol dépassant la limite fixée par le CIO et la présence de composés malodorants (dont l’éthanol) ont été observés lors de récoltes plus tardives. En outre, deux facteurs technologiques, notamment un stockage inadéquat des fruits et de mauvaises pratiques hygiéniques de fabrication, ont favorisé la lipolyse enzymatique du triacylglycérol du fruit modifiant de manière significative les profils d'arômes et d’acides gras de l'huile d'olive.La 3D-FFFS et l'Ultra-FGC ont toutes deux montré de très bonnes performances. La 3D-FFFS couplée à des techniques chimiométriques a été appliquée sur des qualités hétérogènes et dégradées d'échantillons d'huile d'olive libanaise afin de prédire les principaux paramètres physicochimiques de qualité. Ainsi, vingt-deux modèles de régression MLR basés sur les scores PARAFAC ont été générés, dont la majorité a montré un bon coefficient de corrélation (R>0,7). Un deuxième modèle, utilisant la PLS sur les matrices d'émission-excitation (EEM) dépliées, a conduit à des résultats similaires, avec une légère amélioration par rapport au modèle MLR. D’autre part, l’Ultra Flash GC a permis d’identifier en quelques minutes seulement (< 2 min) l'éthanol, le (E,E)-2,4-décadiénal (défaut organoleptique) et le 1-hexanol (fruité, herbeux) comme principaux volatils caractérisant la variété Soury.Cette étude offre la possibilité d’établir au Liban un plan de contrôle analytique qui lie les aspects environnementaux et les techniques de culture/récolte aux caractéristiques physico-chimiques de l'huile d'olive qui en résultent. Une telle matrice monitorée à l’aide de techniques d'analyse rapide facilitera la vérification de la conformité du produit final aux normes internationales. En outre, ce travail préparera le terrain grâce à une fiche d'identification détaillée pour l'IGP
Olive oil is a vital component of the Mediterranean diet, hence Lebanese, owed to its well-known economic and nutritional value. Several environmental, agricultural, and technological factors play an essential role in defining olive oil's quality. In Lebanon, preliminary studies on the quality of extra virgin olive oil have shown that certain quality criteria exceed the International Olive Council's (IOC) standards. However, the causes of such non-conformities have not been clearly identified. Accordingly, ninety-six olive oil samples have been harvested from two seasons, processed using different extraction methods, and collected from eight locations (Akkar, Chouf, Hasbaya, Koura, Tyr, Nabatiyeh, Zgharta, and Hermel). These locations are identified by the European Union to have potentials for Protected Geographical Indications (PGI). In this perspective, and to meet the European framework's requirements, the analyzed oil will be subjected to conventional chemical analysis as suggested by the IOC and to ultra-fast analysis using 3D-front face spectroscopy (3D-FFFS) and ultra-flash gas chromatography (Ultra-FGC).A correlation between the fatty acid profile and the pedoclimatic conditions of the main olive growing regions in Lebanon was noticed. Three main pedoclimatic conditions, altitude, temperature, and relative humidity, were the major influencers and the reason for the distinctive fatty acid profile of the Lebanese olive oil. Lebanese areas with high altitudes, low average temperature, and low relative humidity have high oleic acid content. As for areas with lower altitudes, higher average temperature, and higher relative humidity, the fatty acid profile was characterized by linoleic, linolenic, palmitoleic, and palmitic acids. In addition to the environmental factors, agricultural ones, particularly the harvest date, had affected the chemical constituents of olive oil. The results obtained showed that the harvest date strongly influenced acidity and total polyphenols. A change in the fatty acid profile characterized by a higher linoleic and lower oleic content, an increase in ∆^7-stigmastenol exceeding the limit set by the IOC standards, and a dominating off-flavor compound (ethanol) was noticed as a result of delaying the harvesting time. Besides, two technological factors, particularly improper fruit storage, and bad hygienic practices, significantly affected olive oil’s quality parameters and fatty acid content.3D-FFFS and Ultra-FGC were used in-line with conventional analysis, and they both showed an undeniable performance. 3D-FFFS coupled with chemometric tools, namely multiple linear regression (MLR) applied on parallel factor (PARAFAC) scores and partial least squares (PLS), was tested on inconsistent qualities of olive oil samples to predict quality parameters. Twenty-two MLR models were generated, the majority of which showed a good correlation coefficient (R>0.7). A second model using PLS on the unfolded emission-excitation matrices was also conducted to improve the regression and assess whether the variability can be handled successfully. However, similar results, with a slight improvement over the MLR model, were obtained. As for Ultra Flash GC, it made it possible to identify, in only a few minutes (< 2 min), ethanol, (E,E)-2,4-decadienal (organoleptic defect), and 1-hexanol (fruity, grassy) as the main volatiles characterizing the Soury variety.This study offers the potential to disseminate an analytical control plan that links environmental aspects in Lebanon and cultivation/harvesting techniques to olive oil's resulting physicochemical characteristics. Such a matrix incorporating rapid analysis techniques will facilitate governance over the end product's final quality and, subsequently, conformity to IOC standards. Furthermore, this work will set the ground through a detailed identification fiche for PGI

L'uva contiene un vasto assortimento di metaboliti primari e secondari, che sono importanti sia per la qualità e le proprietà sensoriali del vino che per l'effetto positivo sulla salute umana.La tesi è dedicata allo studio del metaboloma della bacca d'uva attraverso LC-MS, GC-MS e analisi di dati multivariata. Il capitolo 1 è un'introduzione sulla metabolomica dell'uva. La prima parte è dedicata all'importanza dei metaboliti dell'uva mentre la seconda riguarda le diverse piattaforme tecnologiche utilizzate nel campo della metabolomica dell'uva e le analisi di dati multivariata, in particolare con metodo OPLS/O2PLS.Nella terza parte sono riportati alcuni trend recenti nel campo della metabolomica. Il capitolo 2 è dedicato allp sviluppo di nuovi approcci metodologici per studiare il terroir,ovvero l'effetto dell'ambiente e pratiche viticolturali sulla qualità dell'uva. I metaboliti secondari e quelli volatili sono stati analizzati in bacche della varietà Corvina raccolti in 3 annate in 11 vigneti nell'area di Verona. Sono stati proposti due diversi approcci. Un'analisi preliminare unsupervised ha evidenziato il forte effetto del clima sul metaboloma di Corvina. L'approccio più corretto è risultato quindi essere l'analisi delle tre annate contemporaneamente. In seguito, diversi modelli multivariata sono stati costruiti e validati per evitare problemi di overfitting. Alcuni modelli, oltre a quello delle annate, sono risultati essere validi e hanno dato informazioni in termini di metaboliti. In particolare, due vigneti localizzati nell'area del Lago di Garda e caratterizzati da suoli franchi, sono risultati essere ricchi di metaboliti volatili della classe dei benzeni rispetto agli altri vigneti. E' stato poi proposto un secondo approccio di tipo unsupervised in cui è stata ridotta la variabilità nel metaboloma di bacche dovuta al clima, per studiare il terroir come il set di metaboliti che non mostra variabilità dovuta all'annata. L'analisi del dataset di LC-MS ha mostrato che specifici livelli di metaboliti secondari caratterizza in modo diverso i vigneti. Dall'analisi della matrice di GC-MS due cluster di produttori sono stati evidenziati ma nessuna caratteristica dei vigneti analizzata nel lavoro è in grado di spiegare il raggruppamento. Specifiche caratteristiche del terroir influenzano quindi il metaboloma di Corvina, ma queste catatteristiche sono correlate a sottili variazioni di un vasto numero di metaboliti più che da poche e importanti variazioni di pochi metaboliti. Il terzo capitolo è dedicato allo studio delle risposte metaboliche di sei differenti varietà d'uva (Cabernet Sauvignon, Corvina,Merlot,Oseleta, Sangiovese e Shiraz) durante il processo di appassimento in fruttai. Un primo approccio è basato sull'analisi dei dati crudi e ha permesso di discriminare le sei varietà in base al loro metaboloma. Un secondo approccio è basato sull'analisi dei dati normalizzati per il calo ponderale in modo da compensare l'effetto di concentrazione dovuto alla perdita d'acqua. La modulazione di antocianine, flavonoids, flavan-3-oli e procianidine, acidi idrossibenzoici e idrossicinnamici sono dovute principalmente all'effetto di concentrazione e alla degradazione. I composti appartenenti alla classe di stilbeni e viniferine sono risultati essere ampiamente indotti durante il processo di appassimento in tutte le varietà eccetto Cabernet Sauvignon; tali metaboliti potrebbero avere un ruolo nella difesa durante le condizioni di stress dell'appassimento. L'ultima parte del capitolo è dedicata all'analisi della modulazione dei metaboliti volatili delle varietà Corvina e Shiraz. I sesquterpeni, una classe di molecole terpeniche, sono ampiamente indotti in Corvina durante l'appassimento e potrebbero avere un ruolo nel signalling cellulare e nella risposta agli stress.
Grape vine is one of the most important fruit crop in the world. Numerous grape varieties are grown worldwide for production of wine or food. Grape berries contain a large and diverse range of primary and secondary metabolites, that are important for wine quality and sensory attributes but also for their positive impact on human health. This thesis is dedicated to the study of grape metabolomes through LC-MS, GC-MS and advanced multivariate data analysis. This thesis is divided into three main chapters, covering different aspects of grape metabolomics. Chapter 1 is a review on grape metabolomics. The first part is dedicated to the importance of grape and its metabolites, with particular regards on the effect on wine sensory properties. The second part covers the different analytical platforms used in grape metabolomics and the importance of multivariate data analysis, particularly with newly developed OPLS/O2PLS. In the third part, some recent trends in grape metabolomics are described. Chapter 2 is dedicated to the development of new methodological approaches to study terroir, or the effect of environmental and viticultural features on grape and wine quality. Secondary non-volatile and volatile metabolites were extracted from berries of cv. Corvina harvested in three vintages (2006, 2007 and 2008) in eleven vineyards located in the area surrounding Verona. Two different approaches were proposed, based on multivariate data analysis. First, a preliminary unsupervised analysis allowed highlighting the strong effect of climate on cv. Corvina metabolome. The most correct approach resulted to be the analysis of the three vintages together. Several multivariate models were built using vineyard features and validated to avoid over fitting. Few models, beside the one of vintages, gave information in terms of metabolites. Particularly, two vineyards located in Lake Garda area and with by loam soils were characterized by high contents of benzene derivative volatiles compared to other vineyards. A second and completely unsupervised approach was used to reduce the variability in the metabolome that was due to climate, so that to study terroir as the set of metabolites that did not show a year-dependent variation. The analysis of LC-MS dataset revealed that each single vineyard was characterized by specific levels of secondary metabolites. When analysing the GC-MS matrix, two clusters of producers where obtained but no single feature of terroir considered in this survey could explain the grouping. Specific terroir features were suggested to affect the metabolome composition, but such features were mainly correlated to very subtle variation of large sets of metabolites rather than to sharp variation of few metabolites. Chapter 3 is dedicated to the study of the metabolic responses of six different grape varieties (Cabernet Sauvignon, Corvina, Merlot, Oseleta, Sangiovese and Shiraz) during postharvest withering process. A first approach was based on the analysis of raw data and allowed the discrimination of the six varieties by their metabolomes. The second approach was based on the analysis of data matrix normalized for the weight loss, to compensate the concentration effect due to water loss. Modulation of anthocyanin, flavonoid, flavan-3-ol and procyanidin, hydroxybenzoic and hydroxycinnamic acid classes were mainly due to concentration effect and degradation. Compounds belonging to class of stilbenes and viniferins were highly induced during postharvest withering in all the varieties but Cabernet Sauvignon and might act as defence mechanism during stress conditions. The last part of Chapter 3 is focused on the different modulation of volatile compounds in Corvina and Shiraz varieties. Sesquiterpenes, a class of terpene molecules, were highly induced in Corvina during withering and may have a role in stress response and signalling.

During the last decade, many baseline estimation methods have been proposed, but many of these approaches are either only useful for specific kinds of analytical signals or require the adjustment of many parameters. This complicates the selection of an appropriate approach for each kind of chemical signal and the optimization of multiple parameters itself is not an easy task. In this work, an asymmetric least squares (ALS) approach is used with truncated and augmented Fourier basis functions to provide a universal basis space for baseline approximation for diverse analytical signals. The proposed method does not require extensive parameter adjustment or prior baseline information. The basis set used to model the baselines includes a Fourier series truncated to low frequency sines and cosines (consistent with the number of channels) which is then augmented with lower frequencies. The number of basis functions employed depends mainly on the frequency characteristics of the baseline, which is the only parameter adjustment required for baseline estimation. The weighting factor for the asymmetric least squares in this case is dependent mainly on the level of the noise. The adjustment of these two parameters can be easily performed by visual inspection of results. To estimate and eliminate the baseline from the analytical signals, a novel algorithm, called Truncated Fourier Asymmetric Least Squares (TFALS) was successfully developed and optimized. It does not require baseline representative signals or extensive parameter adjustments. The method is described only with parameters optimization using simulated signals. The results with simulated and experimental data sets having different baseline artefacts show that TFALS is a versatile, effective and easy-to-use baseline removal method.