Submitted by Renata Lopes (renatasil82@gmail.com) on 2016-05-02T18:29:34Z No. of bitstreams: 1 leticiastephantavares.pdf: 3190443 bytes, checksum: 26aa58db0ef5c5e3bfe0caaaf2d8671e (MD5) / Approved for entry into archive by Adriana Oliveira (adriana.oliveira@ufjf.edu.br) on 2016-06-03T15:35:59Z (GMT) No. of bitstreams: 1 leticiastephantavares.pdf: 3190443 bytes, checksum: 26aa58db0ef5c5e3bfe0caaaf2d8671e (MD5) / Made available in DSpace on 2016-06-03T15:35:59Z (GMT). No. of bitstreams: 1 leticiastephantavares.pdf: 3190443 bytes, checksum: 26aa58db0ef5c5e3bfe0caaaf2d8671e (MD5) Previous issue date: 2015-04-29 / CAPES - Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / O uso incorreto de antibióticos vem se tornando, nos últimos anos, um grande problema de acordo Organização Mundial da Saúde, uma vez que tem aumentado o número de microrganismos resistentes aos medicamentos mais frequentemente utilizados. Em contrapartida a este problema de saúde pública, novos antibióticos com diferentes vias de ação têm sido pesquisados. Muitos destes antimicrobianos têm sido descobertos a partir da primeira linha de defesa de vegetais e de animais. Estas moléculas são denominadas Peptídeos Antimicrobianos (AMPs). No intuito de encontrar novos compostos com atividade antimicrobiana foram desenvolvidos 3 peptídeos com ação bactericida a partir do transcriptoma de Lippia rotundifolia e L. alba. O RNA normalizado foi sequenciado utilizando-se a plataforma 454 GS e a partir do mesmo foram gerados e modelados in silico peptídeos com estrutura e ação semelhantes a AMPs. Em seguida os peptídeos foram sintetizados e sua atividade validada por testes antimicrobianos. Os peptídeos Lalb1 e Lrot3 apresentaram resultados promissores e foram remodelados a partir de um desenho racional visando obter a melhor estrutura e atividade dos mesmos. Os peptídeos L.rot3.5 e L.rot3.6 apresentaram os resultados mais promissores contra os patógenos testados. Os resultados aqui demonstrados sugerem que o uso de transcriptomas é uma importante ferramenta para a descoberta de novos AMPs com ação contra bactérias Gram-positivas e Gram-negativas. / Misuse of antibiotics has become, in recent years, worldwide problem according to the World Health Organization, since the number of resistant microorganisms for most commonly used drugs has increased. In contrast to this public health problem, new antibiotics with different courses of action have been researched. Many of these antibiotics have been discovered from the first line of plant and animal defense. This is a group of molecules called Antimicrobial Peptides (AMPs). In the present work three antimicrobial peptides showing bactericidal activity were developed from the transcriptome of Lippia rotundifolia and L. alba. The normalized RNA was sequenced in 454 GS platform and the RNA library was used for in silico searching and modeling peptides showing similar structure and action to AMP. The peptides were synthesized and their activity was validated by antimicrobial tests. The L.alb1 and L.rot3 peptides showed promising results and were modelled again by the use of rational design methodology to inbreed structure and activity. The L.rot3.5 and L.rot3.6 peptides showed the most promising results against the tested pathogens. The results reported here demonstrated that the discovery of new AMPs from transcriptome against Gram-positive and Gram-negative bacteria is an important tool for this purpose.

Subsurface microbes are extremophiles adapted to thrive in deep, resource-limited environments, performing crucial roles in a myriad of biogeochemical processes. The extremozymes they produce might play a pivotal role in catalyzing these processes. Identifying and characterizing those enzymes could contribute to the advancements in industrially important biocatalytic reactions. Among various enzymes, sorbitol dehydrogenases are enzymes that catalyze the reversible conversion of sorbitol into fructose in the presence of NAD+. In this study, we focus on the exploration of a sorbitol dehydrogenase (SDHSIUC1) derived from the novel strictly anaerobic, thermophilic, subsurface bacterium, Caldiatribacterium inferamans SIUC1, which is one of the first cultured members from the candidate phylum Atribacteria OP9. As SDHSIUC1 originated from a subsurface microbe, we hypothesized that the enzyme has industrially beneficial characteristics such as higher thermostability and can be used for bioindustry applications such as synthesis of rare sugars and chiral alcohols. We successfully cloned, expressed, and purified the functional SDHSIUC1 enzyme aerobically using E. coli BL21(DE3) and did biochemical assays to characterize its properties. Additionally, in combination with the findings of biochemical characterization, we applied in silico approaches such as molecular modeling and molecular docking to describe the functional mechanism of the enzyme. Initial phylogenetic tree analysis using a pool of 24 amino acid sequences showed that the closest relative for SDHSIUC1 is a Candidatus Caldiatribacterium californiense, which is an uncultured member of the Atribacteria phylum. Size exclusion chromatography and Native-PAGE suggested that SDHSIUC1 is a hexamer with a size of 225 kDa. Kinetic characterization of the SDHSIUC1 showed that the enzyme has a higher affinity for sorbitol and fructose in the presence of NAD+ and NADH, respectively. Furthermore, SDHSIUC1 enzyme is promiscuous as it could utilize other polyols (i.e., glycerol, xylitol, inositol), diols (i.e., butanediol), aldehydes (i.e., glycolaldehyde), and ketoses (i.e., sorbose) in the presence of NAD+/NADH cofactors. We observed a significant increase in enzyme activity in the presence of Zn2+, where other metal ions such as Mn2+ and Mg2+ also resulted in rate improvements. The enzyme is an alkaline dehydrogenase that prefers a higher pH above 8. The effect of temperature on SDHSIUC1 activity showed that it’s a thermophilic enzyme with activity at 85 ℃. The thermal denaturation points of the enzyme at 85 ℃ was increased when the enzyme was preincubated at 85 ℃ in the presence of Zn2+. Notably, the enzyme preincubated 25 min at 85 ℃ in the presence of Zn2+ prefers fructose conversion and ceased the sorbitol conversion. We identified the presence of a structural Zn2+ binding site in SDHSIUC1 in addition to a catalytic Zn2+ binding site. We speculated that the structural Zn2+ involves thermal stability of the enzyme. Hence, we mutated the cysteine with serine of potential structural Zn2+ binding site (Cys96, Cys99, Cys102, and Cys110). Indeed, the Inductively coupled plasma mass spectrometry (ICP-MS) analysis revealed the mutated enzyme contains a lower amount of Zn2+ relative to the native enzyme. The data revealed that the mutated enzyme has low melting temperature (78 ℃) relative to the native enzyme (92 ℃), suggesting that structural Zn2+ is key to enhance the thermal stability of the SDHSIUC1. Surprisingly, we observed that the mutant enzyme completely lost its activity. The data suggests the role of structural Zn2+ binding site on both the structural and functional stability of SDHSIUC1. In consistence with the in-vitro data, the preliminary computational modeling data suggest that the losing structural Zn2+ unstable the enzyme and we are conducting in depth in-silico study to unveil the mechanism(s). We envisioned that the mechanisms behind the thermal stability of SDHSIUC1 could be used as basic model to enhance thermostable protein for the industrial application (e.g., design thermostable plastic hydrolyzing enzymes). To further demonstrate the potential applications of the SDHSIUC1, we genome-integrated it into the industrially important microorganism Pseudomonas putida KT2440. The resulting strain exhibited significantly increased growth in the presence of sorbitol compared to the wild-type P. putida KT2440, highlighting the potential of this enzyme for industrial applications such as enabling sorbitol catabolism or establishing xylose reductase pathway in P. putida KT2440 (i.e., leverage xylitol dehydrogenase activity of SDHSIUC1). In summary, this study has uncovered a novel thermostable sorbitol dehydrogenase from a subsurface microbe, which could have potential applications in the bioindustry where thermostable sorbitol dehydrogenases are required for the application in food and beverage industry, pharmaceutical industry, biofuel production etc. as it would be advantageous for the industrial processes.

Sepsis is defined as life-threatening organ dysfunction caused by the body’s dysregulated host response to an infection. An early feature of sepsis is the dysregulated activation of endothelial cells; this process initiates a cascade of inflammatory events by releasing mediators, leading to leukocyte (e.g., neutrophil) adhesion, migration, tissue damage and multiple organ dysfunction syndrome. Therapeutic approaches for the treatment of sepsis are largely supportive, but there are no specific therapeutics that sustain neutrophil-endothelium function. All sepsis drugs developed in mice models have failed in clinical trials. Thus, novel methods leveraging recent developments in omics are needed to evaluate the neutrophil-endothelium response to therapeutics in sepsis. In this work, bioinformatics, proteomics and network modeling tools were employed to investigate proteomics in organ-specific endothelial cells (e.g., lungs, liver and kidneys) under inflammatory conditions and in neutrophils from sepsis patients. I hypothesized that the network models and the subsequent identification of drug targets for the repurposing of therapeutics will provide novel insight on the role of proteins in predicting pathological responses in endothelial cells and neutrophils under inflammatory/septic conditions. In Aim 1, using in vitro and in silico analysis, mouse lung, liver and kidney endothelial cells were exposed to inflammatory conditions to mimic sepsis over time to investigate endothelial cell differential protein expression and the underlying mechanisms contributing to inflammation. Critical findings included the lung having the highest number of differentially expressed proteins across time, processes such as cell adhesion, apoptosis and angiogenesis impacting organs (other than defense and immune processes) and a uniformity in protein expression being observed across time. In Aim 2, neutrophil immunophenotyping was performed using our Organ-on-Chip to identify three functional phenotypes (Hyperimmune, Hypoimmune and Hybrid) based on ex vivo adhesion and migration patterns on endothelial cells. Proteomics was used to identify unique signatures that correlated with disease severity; for example, two phenotypes (Hyperimmune and Hybrid) had similar protein expression across different protein classes (e.g., higher protein expression in neutrophil adhesion, cytoskeleton and defense classes) compared to the Hypoimmune phenotype. This novel approach could advance precision medicine for sepsis patients. In Aim 3, not only was functional enrichment analysis performed to identify critical processes/pathways in the neutrophil functional phenotypes from Aim 2, but in silico modeling was used to identify Food and Drug Administration approved drugs for treating sepsis. Specifically, network model(s) were built to investigate protein connectivity between those proteins that are targeted by Food and Drug Administration approved drugs and those that are potential druggable candidates. Furthermore, a degree-centric approach was used to rank the proteins in the network models to determine their role in regulating network functionality in the different processes/pathways associated with sepsis. Proteins that were highly ranked in each phenotype included Isocitrate Dehydrogenase (NAD(+)) 3 Catalytic Subunit Alpha (IDH3A - Hybrid), Cytochrome B5 Reductase (CYB5R3 – Hypoimmune) and Palmitoyl-Protein Thioesterase (PPT1 - Hyperimmune phenotype). Drugs targeting the highly ranked proteins included Palmitic acid (targeting PPT1 in Hyperimmune), NADH and FAD (targeting CYB5R3 in Hypoimmune) and NADH, Copper and Manganese (targeting IDH3A in Hybrid). Thus, these proteins should be further validated experimentally in each phenotype using our Organ-on-Chip. / Bioengineering

In silico experimentation is the process of using computational and statistical models to predict medicinal properties in chemicals; as a means of reducing lab work and increasing success rate this process has become an important part of modern drug development. There are various ways of representing molecules - the problem that motivated this paper derives from collecting substructures of the chemical into what is known as fractional representations. Assembling large sets of molecules represented in this way will result in sparse data, where a large portion of the set is null values. This consumes an excessive amount of computer memory which inhibits the size of data sets that can be used when constructing predictive models.In this study, we suggest a set of criteria for evaluation of random forest implementations to be used for in silico predictive modeling on sparse data sets, with regard to computer memory usage, model construction time and predictive accuracy.A novel random forest system was implemented to meet the suggested criteria, and experiments were made to compare our implementation to existing machine learning algorithms to establish our implementation‟s correctness. Experimental results show that our random forest implementation can create accurate prediction models on sparse datasets, with lower memory usage overhead than implementations using a common matrix representation, and in less time than existing random forest implementations evaluated against. We highlight design choices made to accommodate for sparse data structures and data sets in the random forest ensemble technique, and therein present potential improvements to feature selection in sparse data sets. / Program: Systemarkitekturutbildningen

Flavonoid glycosyltransferases (GTs), enzymes integral to plant ecological responses and human pharmacology, necessitate rigorous structural elucidation to decipher their mechanistic function and substrate specificity, particularly given their role in the biotransformation of diverse pharmacological agents and natural products. This investigation delved into a comprehensive exploration of the flavonol 3-O GT from Citrus paradisi (Cp3GT), scrutinizing the impact of a c-terminal c-myc/6x histidine tag on its enzymatic activity and substrate specificity, and successfully achieving its purification to apparent homogeneity. This established a strong foundation for potential future crystallographic and other structure/function analyses. Through the strategic implementation of site-directed mutagenesis, a thrombin cleavage site was incorporated proximal to the tag, followed by cloning in Pichia pastoris, methanol-induced expression, and cobalt-affinity chromatography for initial purification stages. Notably, the recombinant tags did not exhibit a discernible influence on Cp3GT kinetics, substrate preference, pH optima, or metal interactions, maintaining its specificity towards flavonols at the 3-OH position and favoring glucosylation of quercetin and kaempferol. Subsequent purification steps, including MonoQ anion exchange and size-exclusion chromatography, yielded Cp3GT with ≥95% homogeneity. In silico molecular models of Cp3GT and its truncated variants, Cp3GTΔ80 and Cp3GTΔ10, were constructed using D-I-TASSER and COFACTOR to assess binding interactions with quercetin and kaempferol. Results indicated minimal interference of c-myc/6x-his tags with the native Cp3GT structure. This study not only lays a foundation for impending crystallographic studies, aiming to solidify the understanding of Cp3GT's stringent 3-O flavonol specificity, but also accentuates the potential of microbial expression platforms and plant metabolic engineering in producing beneficial compounds. To this end, a thorough review of four pivotal classes of plant secondary metabolites, flavonoids, alkaloids, betalains, and glucosinolates, was conducted. This will open avenues for further research and applications in biotechnological, medical, and agricultural domains.

No / Review article. No abstract.

L’objectif de ce projet de thèse est d’étudier et caractériser les mécanismes impliqués dans la bascule respiro-fermentaire chez des cellules eucaryotes dotées d’un métabolisme mitochondrial. Les cellules eucaryotes ont des besoins différents en oxygène pour la production d’énergie et leur survie dans un environnement donnée. Elles sont qualifiées de type aérobie stricte lorsque la présence d’oxygène leur est nécessaire ou aéro-anaérobie facultatif dans le cas où l’oxygène n’est pas indispensable à la production d’énergie. La levure Yarrowia lipolytica a été choisie comme modèle d’étude de par sa particularité à être un micro-organisme aérobie strict avec une grande capacité d’accumuler de lipides et de production d’acides organiques. Les études expérimentales et analytiques, par l’emploi de méthodes mathématiques de modélisation du métabolisme, ont permis d’identifier des contraintes métaboliques impliquées dans la transition respiro-fermentaire chez cette levure au métabolisme énergétique oxydatif. La production de l’acide citrique par Y. lipolytica, déjà rapportée dans la littérature, a été choisi comme un marqueur de cette transition respiro-fermentaire. Nous avons découvert que l’inhibition de la protéine oxydase alternative (AOX), impliquée dans la respiration mitochondriale, par la molécule n-propyl gallate (nPG) permet d’améliorer le rendement de production d’acide citrique par fermentation du glucose dans une culture de Y. lipolytica. Ces résultats montrent que la nPG, déjà utilisée dans l’industrie agro-alimentaire et pharmaceutique en tant que conservateur joue sur la bascule respiro-fermentaire par inhibition de la consommation d’oxygène et stimule ainsi la production d’acide citrique. La modélisation du réseau métabolique de Y. lipolytica, décrit à l’échelle du genome, par dynamic Flux Balance Analysis (dFBA) a permis d’identifier l’accumulation des espèces oxydantes dites ROS (Reactive Oxygen Species) comme un levier majeur de la bascule respiro-fermentaire et donc de la production d’acide citrique chez la levure Y. lipolytica. De plus, nos résultats préliminaires montrent que l’oxydation des lipides accumulés par Y. lipolytica pourrait être à l’origine de la génération des ROS. Cette étude doit être approfondie expérimentalement et constitue un apport important pour l’industrie agro-alimentaire et pharmaceutique.Mots clés : Bascule respiro-fermentaire, Acide citrique, lipides, Yarrowia lipolytica, n-propyl gallate, Reactive Oxygen Species, modélisation, dynamic Flux Balance Analysis / The main goal of this thesis project is to study and characterize mechanisms involved in respiratory to fermentative shift in eukaryotic cells endowed with mitochondrial metabolism. Eukaryotic cells have different oxygen requirements for energy production and survival in a given environment. They are described as strict aerobic when the presence of oxygen is necessary or optional aero-anaerobic in when oxygen is not essential for energy production. The yeast Yarrowia lipolytica was chosen as our study model thanks to its particularity since it is a strict aerobic microorganism with a high capacity to accumulate lipids and to produce organic acids. Experimental and analytical studies, using mathematical methods for modeling cell metabolism, allowed us to identify metabolic constraints involved in respiratory to fermentative transition in this yeast showing oxidative energy metabolism. Production of citric acid by Y. lipolytica, already reported in the literature, has been chosen as a marker for this in respiratory to fermentative shift. We found that the inhibition of the alternative oxidase protein (AOX) involved in mitochondrial respiration, by adding n-Propyl gallate (nPG) molecule improves the yield of citric acid production by fermentation of glucose in a Y. lipolytica culture. These results show that nPG, already used in food and pharmaceutical industry as a preservative, plays on respiratory to fermentative balance by inhibition of oxygen consumption and thus stimulates the production of citric acid. Modeling of the metabolic network of Y. lipolytica, described at genome-scale, by dynamic Flux Balance Analysis (FBA) has identified the accumulation of intracellular ROS (Reactive Oxygen Species) species as major levers for respiratory to fermentative shift and therefore the production of citric acid by Y. lipolytica. Therefore, our preliminary results show that oxidation of lipids accumulated by Y. lipolytica could be involved in generation of ROS species. This study must be experimentally deepened and constitutes an important contribution for the agri-food and pharmaceutical industry.Key words: Respiratory to fermentative shift, Citric acid, lipids, Yarrowia lipolytica, n-Propyl gallate, Reactive Oxygen Species, modeling, dynamic Flux Balance Analysis

Dans les nouveaux protocoles thérapeutiques par voie inhalée, le dosage des particules actives reste un problème complexe qui dépend de trois principaux facteurs : leur taille, la dynamique des flux et les variations de calibre bronchique. La solution nécessite de disposer d'un modèle de distribution des gaz et aérosols administrés dans les poumons. Ventilation pulmonaire et effets du cycle respiratoire sur la dynamique des fluides deviennent deux enjeux clés de la pratique clinique.<br /><br />Dans ce contexte, le projet RNTS RMOD a pour objectif de développer un simulateur morpho-fonctionnel des voies respiratoires pour l'aide au diagnostic, au geste médico-chirurgical et à l'administration de médicaments par inhalation.<br /><br />Contribuant au projet RMOD, la recherche développée dans cette thèse propose une modélisation in-silico de la structure des voies aériennes supérieures (VAS) et proximales (VAP) à partir d'examens tomodensitométriques (TDM). L'investigation morphologique et la simulation fonctionnelle bénéficient alors de géométries 3D réelles, adaptées au patient et spécifiques des pathologies rencontrées.<br /><br />La modélisation développée fait coopérer des méthodes originales de segmentation, de construction de surface maillée et d'analyse morpho-fonctionnelle.<br /><br />La segmentation des VAP est obtenue par un schéma diffusif et agrégatif gouverné par un modèle markovien, dont l'initialisation repose sur l'opérateur de coût de connexion sous contrainte topographique. De cette segmentation, l'axe central de l'arbre bronchique est extrait de manière robuste et précise en combinant information de distance, propagation de fronts, et partition conditionnelle locale. Cet axe central est représenté sous forme d'une structure hiérarchique multivaluée synthétisant caractéristiques topologiques et géométriques de l'arbre bronchique. Une surface maillée est ensuite construite en appliquant une procédure de Marching Cubes adaptative, les paramètres des différents filtres mis en jeu étant automatiquement ajustés aux caractéristiques locales du réseau bronchique conditionnellement aux attributs de l'axe central.<br /><br />La segmentation des VAS repose sur une propagation markovienne exploitant les variations locales de densité. L'initialisation combine morphologie mathématique et information de contour afin de garantir la robustesse à la topologie. Une procédure de type triangulation de Delaunay restreinte à une surface fournit ensuite la représentation maillée des VAS. Il est établi que la topologie et la géométrie des structures complexes composant les VAS sont effectivement préservées.<br /><br />Pour permettre aux médecins de valider les modèles maillés ainsi construits, un environnement virtuel 3D convivial et interactif a été réalisé. En outre, la morphologie des voies aériennes exo- et endo-luminale est analysée de façon automatique à partir de simulations d'écoulement pour des géométries réelles.<br /><br />Enfin, une modélisation unifiée des VAP et VAS est obtenue pour la première fois. Elle démontre la pertinence des approches développées. Elle ouvre la voie à la construction de modèles in-silico complets de l'appareil respiratoire ainsi qu'aux études fonctionnelles prenant en compte les paramètres morphologiques susceptibles d'influer localement ou globalement sur la dynamique des écoulements.

<p>New effective experimental techniques in medicinal chemistry and pharmacology have resulted in a vast increase in the number of pharmacologically interesting compounds. However, the number of new drugs undergoing clinical trial has not augmented at the same pace, which in part has been attributed to poor absorption of the compounds.</p><p>The main objective of this thesis was to investigate whether computer-based models devised from calculated molecular descriptors can be used to predict aqueous drug solubility, an important property influencing the absorption process. For this purpose, both experimental and computational studies were performed. A new small-scale shake flask method for experimental solubility determination of crystalline compounds was devised. This method was used to experimentally determine solubility values used for the computational model development and to investigate the pH-dependent solubility of drugs. In the computer-based studies, rapidly calculated molecular descriptors were used to predict aqueous solubility and the melting point, a solid state characteristic of importance for the solubility. To predict the absorption process, drug permeability across the intestinal epithelium was also modeled.</p><p>The results show that high quality solubility data of crystalline compounds can be obtained by the small-scale shake flask method in a microtiter plate format. The experimentally determined pH-dependent solubility profiles deviated largely from the profiles predicted by a traditionally used relationship, highlighting the risk of data extrapolation. The <i>in silico</i> solubility models identified the non-polar surface area and partitioned total surface areas as potential new molecular descriptors for solubility. General solubility models of high accuracy were obtained when combining the surface area descriptors with descriptors for electron distribution, connectivity, flexibility and polarity. The used descriptors proved to be related to the solvation of the molecule rather than to solid state properties. The surface area descriptors were also valid for permeability predictions, and the use of the solubility and permeability models in concert resulted in an excellent theoretical absorption classification. To summarize, the experimental and computational models devised in this thesis are improved absorption screening tools applicable to the lead optimization in the drug discovery process. </p>

New effective experimental techniques in medicinal chemistry and pharmacology have resulted in a vast increase in the number of pharmacologically interesting compounds. However, the number of new drugs undergoing clinical trial has not augmented at the same pace, which in part has been attributed to poor absorption of the compounds. The main objective of this thesis was to investigate whether computer-based models devised from calculated molecular descriptors can be used to predict aqueous drug solubility, an important property influencing the absorption process. For this purpose, both experimental and computational studies were performed. A new small-scale shake flask method for experimental solubility determination of crystalline compounds was devised. This method was used to experimentally determine solubility values used for the computational model development and to investigate the pH-dependent solubility of drugs. In the computer-based studies, rapidly calculated molecular descriptors were used to predict aqueous solubility and the melting point, a solid state characteristic of importance for the solubility. To predict the absorption process, drug permeability across the intestinal epithelium was also modeled. The results show that high quality solubility data of crystalline compounds can be obtained by the small-scale shake flask method in a microtiter plate format. The experimentally determined pH-dependent solubility profiles deviated largely from the profiles predicted by a traditionally used relationship, highlighting the risk of data extrapolation. The in silico solubility models identified the non-polar surface area and partitioned total surface areas as potential new molecular descriptors for solubility. General solubility models of high accuracy were obtained when combining the surface area descriptors with descriptors for electron distribution, connectivity, flexibility and polarity. The used descriptors proved to be related to the solvation of the molecule rather than to solid state properties. The surface area descriptors were also valid for permeability predictions, and the use of the solubility and permeability models in concert resulted in an excellent theoretical absorption classification. To summarize, the experimental and computational models devised in this thesis are improved absorption screening tools applicable to the lead optimization in the drug discovery process.