280 research outputs found
Pescando conhecimento pelo saber tradicional: reconhecimento individual e aspectos etnoecológicos do boto-da-tainha, Tursiops truncatus, por pescadores artesanais de Laguna, Sul do Brasil
No full textTCC(graduação) - Universidade Federal de Santa Catarina. Centro de Ciências Biológicas. Biologia.O boto-da-tainha, Tursiops truncatus, ocorre em toda a costa sul do Brasil, com destaque para a população residente nas águas próximas ao município de Laguna (SC), que pesca “cooperativamente” com pescadores artesanais. Este trabalho busca descrever e analisar a percepção dos pescadores quanto ao boto-da-tainha e quanto a este evento de pesca cooperativa. Assim, entrevistas estruturadas foram realizadas para: descrever o perfil dos pescadores que interagem com o boto; testar a hipótese de que os pescadores reconhecem os botos individualmente, descrevendo como se dá este processo; coletar informações sobre nome dos botos, sexo e padrões comportamentais; e identificar possíveis ameaças para a conservação desta população. As entrevistas se deram através da aplicação de questionários, com prévia autorização dos entrevistados. Dos 38 entrevistados, 60% possuem idade acima dos 50 anos e participam da pesca cooperativa há mais de 20, evidenciando a importância desta atividade como tradição local. A maioria dos pescadores aprendeu a arte da pesca com os pais, sinalizando uma típica transmissão vertical do conhecimento. Em geral, pescadores que aprenderam a pescar sozinhos discordam ao reconhecer os botos, o que pode indicar um menor conhecimento adquirido. Todos os entrevistados reconheceram pelo menos uma das fotos que lhes foi apresentada sendo que apenas 6,7% delas nunca foram reconhecidas. Aparentemente, botos que possuem marcas de longa duração ou doenças epidérmicas são mais facilmente reconhecidos pelos pescadores. O tipo de marca na nadadeira dorsal também foi determinante no que diz respeito à concordância no reconhecimento entre os pescadores, com relação às fotos apresentadas. Apenas uma das fotografias apresentou concordância total, sendo reconhecida todas as vezes pelo nome de “Caroba”. Outros dois botos também apresentaram valores altos de concordância. Botos considerados “mais velhos”, segundo os pescadores, e que participam da pesca cooperativa, são identificados com maior grau de certeza. A rede de bagre é uma arte de pesca proibida na região e de grande conflito local, sendo apontada pela maioria dos entrevistados (70%) como a principal causa de morte de botos na área. Entre as sugestões apresentadas pelos pescadores para a proteção dos botos, destaca-se a proibição e uma fiscalização mais efetiva com relação à pesca do bagre. Este trabalho também compara as informações científicas e tradicionais e mostra como estas últimas são de fundamental importância na complementação e entendimento do cenário vi local e na construção de uma proposta de manejo e/ou conservação do boto e de seu ecossistema
On Quality Thresholds for the Clustering of Molecular Structures
No full text[Image: see text] It has been recently suggested that diametral (so-called quality) similarity thresholds are superior to radial ones for the clustering of molecular three-dimensional structures (González-Alemán et al., 2020). The argument has been made for two clustering algorithms available in various software packages for the analysis of molecular structures from ensembles generated by computer simulations, attributed to Daura et al. (1999) (radial threshold) and Heyer et al. (1999) (diametral threshold). Here, we compare these two algorithms using the root-mean-squared difference (rmsd) between the Cartesian coordinates of selected atoms as pairwise similarity metric. We discuss formally the relation between these two methods and illustrate their behavior with two examples, a set of points in two dimensions and the coordinates of the tau polypeptide along a trajectory extracted from a replica-exchange molecular-dynamics simulation (Shea and Levine, 2016). We show that the two methods produce equally sized clusters as long as adequate choices are made for the respective thresholds. The real issue is not whether the threshold is radial or diametral but how to choose in either case a threshold value that is physically meaningful. We will argue that, when clustering molecular structures with the rmsd as a metric, the simplest best guess for a threshold is actually radial in nature
On Quality Thresholds for the Clustering of Molecular Structures
No full textIt has been recently
suggested that diametral (so-called quality)
similarity thresholds are superior to radial ones for the clustering
of molecular three-dimensional structures (González-Alemán
et al., 2020). The argument has been made for two clustering algorithms
available in various software packages for the analysis of molecular
structures from ensembles generated by computer simulations, attributed
to Daura et al. (1999) (radial threshold) and Heyer et al. (1999)
(diametral threshold). Here, we compare these two algorithms using
the root-mean-squared difference (rmsd) between the Cartesian coordinates
of selected atoms as pairwise similarity metric. We discuss formally
the relation between these two methods and illustrate their behavior
with two examples, a set of points in two dimensions and the coordinates
of the tau polypeptide along a trajectory extracted from a replica-exchange
molecular-dynamics simulation (Shea and Levine, 2016). We show that
the two methods produce equally sized clusters as long as adequate
choices are made for the respective thresholds. The real issue is
not whether the threshold is radial or diametral but how to choose
in either case a threshold value that is physically meaningful. We
will argue that, when clustering molecular structures with the rmsd
as a metric, the simplest best guess for a threshold is actually radial
in nature
Structure and evolution of protein allosteric sites
No full textLa presente tesis estudia los sitios alostéricos desde una perspectiva estructural y evolutiva. La regulación alostérica es un aspecto fundamental de la vida a nivel molecular, ya que es el mecanismo más potente y frecuente en la regulación de la actividad proteica: mediante la unión de un ligando a un sitio que no es el sitio activo. Este fenómeno fue descrito por primera vez hace más de 50 años y desde entonces no ha dejado de captar la atención de la comunidad científica, llegando incluso a ser calificado como ̀el segundo secreto de la vida', después del código genético. Sin embargo, la comprensión cabal de los mecanismos involucrados continúa siendo un gran desafío científico. Actualmente, los sitios alostéricos han despertado un creciente interés por parte de expertos en química medicinal y compañías farmacéuticas, dado su potencial para el desarrollo de nuevos fármacos. La tesis se presenta como un ̀compendio de publicaciones'. El primer artículo fue publicado a principios del año 2010 y describe la primera etapa del proyecto, una serie de análisis computacionales a gran escala para caracterizar sitios de unión a ligando integrando información referente a secuencia, sitios activos y estructura para más de mil familias proteicas. Mediante la identificación de sitios de unión comunes en distintas estructuras de la misma familia proteica, se desarrolló un método para medir la conservación estructural de dichos sitios. Esta metodología permitió realizar una caracterización de sitios de unión considerando distintos aspectos, como la conservación evolutiva a nivel de secuencia, flexibilidad estructural, potencial electrostático y conservación estructural. El descubrimiento más significativo fue la inesperada falta de correlación entre las medidas de conservación de secuencia y estructura para muchos de los sitios de unión predichos. Este hallazgo es válido también para casos específicos de proteínas alostéricas, donde el sitio activo está conservado tanto a nivel de secuencia como de estructura, pero el sitio alostérico sólo presenta conservación a nivel estructural y no de secuencia. El segundo artículo fue publicado a fines del año 2012 y explora la relación entre la flexibilidad proteica y la regulación alostérica, definiendo una metodología computacional para la predicción de sitios alostéricos en estructuras proteicas. Más allá de los aspectos dinámicos que fueron estudiados mediante el análisis de modos normales, el método también incorpora la medida de conservación estructural desarrollada en el primer artículo. El sistema predictivo fue puesto a prueba utilizando un extenso conjunto de proteínas alostéricas de estructura conocida, obteniendo un valor predictivo positivo de 65%. Después de la segunda publicación, el método se ha implementado como servidor web para brindar apoyo a la investigación de la regulación alostérica, tanto para extender el conocimiento de esta forma fundamental de regulación de la actividad proteica, como para ayudar en la aplicación de dichos conocimientos al desarrollo de nuevos fármacos con objetivos terapéuticos.This thesis studies protein allosteric sites from a structural and evolutionary perspective. Allostery is a fundamental aspect of life at the molecular level, the most common and powerful mechanism of protein activity regulation: through binding of a ligand to a site which is not the active site. This phenomenon was first described more than 50 years ago and it still captures the attention of researchers, while fully understanding its mechanisms remains a grand scientific challenge. Furthermore, allosteric sites have been increasingly calling the attention of medicinal chemists and pharmaceutical companies, given their potential for the development of novel therapeutics. The thesis is presented as a 'compendium of published articles'. The first article was published at the beginning of 2010 describing the first stage of the thesis, a series of large-scale computational analyses to characterize putative small-molecule binding sites by integrating publicly available information on protein sequences, structures and active sites for more than a thousand protein families. By identifying common pockets across different structures of the same protein family a method was developed to measure the pocket's structural conservation. Characterization of putative ligand-binding sites followed using different measures such as sequence conservation, structural flexibility, electrostatic potential and structural conservation. The most relevant finding was the unexpected lack of correlation between the two conservation measures, of sequence and structure, for many of the predicted cavities. This general finding was also observed in specific cases of allosteric proteins, where the active site was conserved both in terms of structure and sequence but the allosteric site was conserved only from the structural perspective and did not show conservation at the sequence level. The second article was published at the end of 2012, it explores the relationship between protein flexibility and allosteric effects defining a computational methodology to predict the presence and location of allosteric sites on protein structures. Besides the dynamical aspects assessed through normal mode analysis, the method also incorporates the structural conservation measure defined in the first article. The predictive approach was benchmarked against a large data set of allosteric proteins of known structure obtaining 65% positive predictive value. After the second publication, the method has been implemented in the form of a freely available web-server aimed to support the work of researchers in the field of allosterism, both to improve the understanding of this fundamental form of protein function regulation and to serve applied purposes in the area of drug design and discovery
A virtual screening procedure combining pharmacophore filtering and molecular docking with the LIE method
No full textActualment, el cribratge virtual juga un paper central en el món del descobriment de fàrmacs. L'anàlisi in silico permet el cribatge de milions de molècules petites i la tria de les més prometedores per a les proves experimentals. Per trobar candidats que puguin esdevenir fàrmacs, és crucial reunir una sèrie d'eines computacionals individuals i complementàries. En aquesta tesi, es descriu un procediment automatitzat de cribatge virtual que combina el modelat de farmacòfors i el seu ús en cerques, mètodes d'alt rendiment d'acoblament molecular, puntuació de consens i estimació d'energia lliure d'unió mitjançant el mètode d'energia d'interacció lineal (LIE) a partir de simulacions de dinàmica molecular. Un dels objectius d'aquesta tesi ha estat el de construir una metodologia flexible i versàtil de cribratge virtual, que permeti la integració de diferents eines en les diferents etapes de l'estudi. El procediment, que es va iniciar com la combinació d'un senzill filtre per tamany, la simulació de l'acoblament molecular i una puntuació de consens, ha derivat en un procediment computacional elaborat i automatitzat amb l'addició de cerques basades en farmacòfor i l'estimació de l'energia lliure d'unió mitjançant el mètode LIE. Aquest mètode integrat té l'objectiu de compensar les debilitats individuals de les diferents tècniques usades i permet avaluar i comparar el rendiment i la l'exactitud d'aquestes tècniques. Una altra fita important ha estat l'aplicació del procediment computacional a proteïnes diana concretes per tal d'avaluar-ne la capacitat de trobar molècules que puguin ser candidats a fàrmacs. Tests experimentals realitzats per a la β-Glucosidasa àcida i la hidrolasa de Bleomicina humanes indiquen que diverses molècules petites seleccionades pel procediment computacional tenen activitat inhibitòria micromolar. El mètode LIE emprat en aquest treball es va aplicar sobre més de deu mil complexos proteïna-lligand per a tres proteïnes diana diferents, el que és, al nostre entendre, la primera aplicació del mètode LIE a aquesta escala.Virtual screening plays a central role in the world of drug discovery today. In silico testing allows to screen millions of small molecules and to choose only the most promising ones for experimental testing. To find potential drug candidates, it is crucial to bring together individual and complementary computational tools. In this thesis, I describe an automated virtual screening procedure that combines pharmacophore modeling and searches, high-throughput molecular docking, consensus scoring and binding free energy estimation with the linear interaction energy (LIE) method through molecular dynamics simulations. One goal of this thesis was to build an evolving and versatile virtual screening methodology, which enables integration of different tools at different steps. The procedure that started as a combination of a simple size filter, molecular docking and consensus scoring, advanced into an elaborate and automated computational workflow with the addition of pharmacophore searches and binding free energy estimation with LIE. This integrated method intends to compensate for weaknesses of individual structure-based techniques and allows the evaluation and comparison of the performance and accuracy of these techniques. Another important goal was to apply the computational workflow to target proteins and find hits that could be drug candidates. Experimental testing performed for human acid β-Glucosidase and bleomycin hydrolase indicate that several small molecules selected by the computational workflow display micromolar inhibitory activity. The standard LIE method used in this work was applied to more than ten thousand ligand-protein complexes for three different targets, which is, to our knowledge, the first time application of LIE at such large scale
Machine Learning techniques in bioinformatics: From data integration to the development of application-oriented tools
No full textEls darrers desenvolupaments en Machine Learning tenen com a objectiu automatitzar els mètodes disponibles, convertint-los en universals alhora que requerint el mínim coneixement expert possible. En aquesta tesi, farem un pas enrere. Ens centrarem en les dades, en les seves necessitats específiques i en com extreure’n informació significativa. Això es farà mitjançant la presentació de diferents treballs que destaquen diversos aspectes a tenir en compte a l’hora de desenvolupar tècniques d’aprenentatge automàtic en bioinformàtica. No hi pot haver cap model sense les consideracions adequades sobre les dades. Per tant, a la primera part, deixarem els models de banda i ens centrarem en la integració de dades. Específicament, presentarem un algorisme per a la normalització de dades de microarrays d’expressió gènica provinents de plataformes distintes. Les dades de microarrays estan àmpliament disponibles als repositoris públics i aquests mètodes permeten la seva posterior anàlisi. A la següent part, considerarem dades de seqüència de pèptids i presentarem una eina per a l'extracció de patrons existents en aquests conjunts. El model, basat en xarxes neuronals convolucionals, és de codi obert i es pot utilitzar per a la predicció d’unió de pèptids a MHC de classe II, entre altres aplicacions. La darrera part es dedicarà a l’anàlisi de dades clíniques. Presentarem un estudi de cohort retrospectiu sobre el càncer de pàncrees. Per a aquest estudi, s’ha desenvolupat una eina per a la predicció de resultats clínicament rellevants.
Des de la integració de dades fins al desenvolupament d’eines orientades a aplicacions, les tres parts que formen aquesta tesi seran autònomes i abordaran diferents reptes en l’àmbit de les aproximacions basades en dades en bioinformàtica.Los desarrollos recientes en Machine Learning tienen como objetivo automatizar los métodos disponibles, haciéndolos universales y requiriendo el menor conocimiento experto posible. En esta tesis, daremos un paso atrás. Nos centraremos en los datos, sus necesidades específicas y cómo extraer información significativa de ellos. Esto se hará a través de la presentación de diferentes trabajos destacando diversos aspectos a considerar a la hora de desarrollar técnicas de Machine Learning en bioinformática. No puede haber ningún modelo sin las consideraciones adecuadas sobre los datos. Por tanto, en la primera parte, dejaremos los modelos a un lado y nos centraremos en la integración de los datos. Específicamente, presentaremos un algoritmo para la normalización de datos de microarrays de expresión génica provenientes de distintas plataformas. Los datos de microarrays están ampliamente disponibles en repositorios públicos y tales métodos permiten su posterior análisis. En la siguiente parte, consideraremos datos de secuencia de péptidos y presentaremos una herramienta para la extracción de patrones existentes en dichos conjuntos. El modelo, basado en redes neuronales convolucionales, es de código abierto y puede ser usado para la predicción de la unión de péptidos a MHC de clase II, entre otras aplicaciones. La última parte estará dedicada al análisis de datos clínicos. Presentaremos un estudio de cohorte retrospectivo sobre cáncer de páncreas. Para este estudio, se ha desarrollado una herramienta para la predicción de resultados clínicamente relevantes.
Desde la integración de datos hasta el desarrollo de herramientas orientadas a aplicaciones, las tres partes que forman esta tesis serán autónomas y cada una abordará diferentes desafíos en el ámbito de las aproximaciones basadas en datos en bioinformática.Recent developments in Machine Learning aim at automatizing available methods, rendering them universal while requiring as little expert-knowledge as possible. In this thesis, we will take a step back. We will focus on the data, their specific needs and how to extract meaningful information out of them. This will be done through the presentation of different works highlighting various aspects to consider when developing Machine Learning techniques in bioinformatics. There cannot be any models without the appropriate considerations on the data. Therefore, in the first part, we will put the models aside and focus on data integration. In more detail, we will present an algorithm for the normalization of gene-expression microarray data across different platforms. Microarray data are widely available in public repositories and such methods enable their subsequent downstream analysis. In the next part, we will consider peptide sequence data and present a tool for the extraction of patterns in such sets. The model, based on convolutional neural networks, is open-source and can be used for peptide MHC-class II binding prediction among other applications. The last part will be dedicated to the analysis of clinical data. We will present a retrospective cohort study on pancreatic cancer. For this study, a tool for the prediction of clinically relevant outcomes has been developed.
From data integration to the development of application-oriented tools, the three parts forming this thesis will be self-contained and will each address different challenges in the realm of data-driven approaches in bioinformatics.Universitat Autònoma de Barcelona. Programa de Doctorat en Bioinformàtic
Prediction of host-pathogen protein-protein interactions: an integrative model
No full textL’ús descontrolat d’antibiòtics contra bacteris patògens ha accelerant l’aparició de bacteris resistents a múltiples fàrmacs en les darreres dècades i s’està convertint en un greu problema de salut pública a nivell mundial. Per combatre aquesta resistència bacteriana, és fonamental obtenir un bon coneixement dels mecanismes d’infecció, que es produeixen en gran mesura per la interacció entre proteïnes del patogen i proteïnes de l’hoste. En aquest sentit, cal millorar l’anotació funcional i la caracterització de les proteïnes implicades i de les seves interaccions. Amb aquest objectiu, hem creat les bases de dades BacFITBase i DualSeqDB, que recopilen la informació disponible sobre la importància de gens bacterians i sobre els canvis d’expressió gènica que es produeixen tant en el patogen com en l’hoste durant el procés d’infecció. Així mateix, hem desenvolupat HPIPred, un sistema de predicció d’interaccions proteïna-proteïna entre hoste i patogen. HPIPred està basat en la codificació numèrica de propietats fisicoquímiques dels aminoàcids i és capaç d’integrar informació fenotípica per guiar el procés de predicció. La combinació d’aquestes eines podria servir com a guia en el desenvolupament de nous fàrmacs contra la resistència bacteriana.El uso descontrolado de antibióticos contra bacterias patógenas está acelerando la aparición de bacterias resistentes a múltiples fármacos y se está convirtiendo en un grave problema en materia de salud pública mundial. Para combatir esta resistencia bacteriana, es fundamental obtener un mayor conocimiento de los mecanismos de infección, que se producen en gran medida por la interacción entre proteínas del patógeno y proteínas del organismo huésped. En este sentido, es necesaria una mejora en la anotación funcional y caracterización de estas proteínas y de sus interacciones. Con este objetivo, hemos creado las bases de datos BacFITBase y DualSeqDB, que recopilan información sobre la importancia de genes bacterianos y sobre los cambios de expresión génica que se producen en el patógeno y en el huésped durante el proceso infeccioso, respectivamente. Del mismo modo, hemos desarrollado HPIPred, un sistema de predicción de interacciones proteína-proteína entre huésped y patógeno, basado en la codificación numérica de propiedades físico-química de los aminoácidos, capaz de integrar información fenotípica para guiar el proceso de predicción. La combinación de estas herramientas podría servir como guía en el desarrollo de nuevos fármacos contra la resistencia bacteriana.The use and misuse of antibiotics against pathogenic bacteria is accelerating the appearance of multi-drug resistant bacteria and it is becoming a serious concern in public health worldwide. In order to fight antibacterial resistance, it is crucial to get a better understanding of the mechanisms underlying infection, which occur to a great extent due to the interaction between host and pathogen proteins. In this sense, it is necessary to improve the functional annotation and characterization of these proteins and their interactions. With this aim, we have created two databases, BacFITBase and DualSeqDB, which gather information on the importance of bacterial genes and the gene expression changes that occur both in the pathogen and in the host during the infection process, respectively. Similarly, we have developed HPIPred, a host-pathogen protein-protein interaction prediction system, based on the numerical encoding of physicochemical properties of amino acids, capable of integrating phenotypic information to guide the prediction process. The combination of these tools could be useful as a guidance in the development of new drugs against antibacterial resistance.Universitat Autònoma de Barcelona. Programa de Doctorat en Bioinformàtic
Algorithms for the Analysis of Biomolecular Simulations: Ensemble Averages, Marginal Distributions, Clustering and Markov Models
No full textThe dynamics of biomolecules, in particular the folding of peptides and proteins, is a highly complex process. The temporal and configurational resolution of experiments is typically too poor to yield a detailed picture of this process.The network of forces which enter the equations of motion governing this process, on the other hand, are too complex to
be analyzed directly. Numerical integration of the equations of motion, however, provides a means to obtain a detailed trajectory of a biomolecule. In particular, numerical integration of the classical equations of motion on an atomic level, known as molecular dynamics (MD), has proved a very powerful tool for the elucidation of the dynamics of biomolecules.
One reason for its success is that MD cannot only be used to generate trajectories but at the same time is an efficient method to sample the equilibrium distribution of the configurations of a molecule. Once the equilibrium distribution is known, most macroscopic properties can be calculated by averaging. But the equilibrium distribution also represents the gate-way to a more detailed understanding of the dynamics of
the molecule. The metastable states of the molecules correspond to regions of high probability density in the configurational space and can, in principle, be extracted directly from the equilibrium distribution. The equilibrium distribution also contains information on how the degrees of freedom of a molecule interact with each other. Analyzing these dependences, one can understand how the conformations of a molecule and ultimately also its dynamics arise from its structural properties.
The development of algorithms used in MD and the tremendous increase in computer power over the last thirty years has generated a need for tools to categorize and concisely represent the enormous amount of data which can be generated by modern MD simulations. Markov models of the dynamics of biomolecules provide such a tool. Using these models the dynamics of the relevant degrees of freedom even of large molecules can be represented by a square matrix with a dimension in the order of 1000. The equilibrium distribution emerges naturally as the first eigenvector of this matrix and metastable states can be extracted simply by grouping states according to their kinetic proximity.
In this thesis methods to analyze the equilibrium distribution of biomolecules are discussed and tested. Chapter 1 shows how the equations of motion for a single system are linked to equations of motion of the probability distribution and how the equilibrium distribution arises from these. We show how thermodynamic properties can be calculated from the equilibrium distribution and discuss the thermodynamics of protein folding. In the last part of chapter 1 we give a short overview of the technical details of molecular simulation and of the historic development of this method.
Chapter 2 treats the construction of stochastic Markov models from deterministic simulations. Emphasis is placed on the assumptions that are made when mapping the equations of motion onto the central quantity of Markov models, the transition matrix. Using a simple two-bit model and simulations of butane, we illustrate in which cases the assumptions are violated and in which cases they are fulfilled. We also review the mathematical properties of transition matrices and discuss their physical interpretation.
Chapter 3 discusses the categorization of an equilibrium distribution in terms of metastable states.The metastable states of a small -peptide are identified using a kinetic cluster algorithm (which is based on a Markov model
of the peptide) and compared to results of geometric cluster algorithms (which are based on a data set which represents the equilibrium distribution). We find that geometric cluster algorithms which use a density estimate as their cluster criterion rather than a cutoff have the best chance of reliably identifying the metastable states of a molecule.
In chapter 4 we establish a link between the equilibrium distribution of the same beta-peptide (and some structurally related peptides) to the features of its atomic configuration. For that we test to which degree the conformations of its dihedral angles depend on each other and show how the marginal distributions of the backbone dihedral angles arise from the atomic configuration of the residues. In particular, we can show that structural features which stabilize the folded conformation in beta-peptides
are not the same as in natural peptides.
Chapter 5 examines the calculation of ensemble averages of properties which depend non-linearly on the configuration of the system on the example of (3)J-coupling constants. When comparing MD data to experimental results, errors in the equilibrium distribution which can be due to force-field or sampling errors, are often corrected by restraining the simulation to the experimental result. This approach is valid as long as the underlying dynamics is not substantially distorted. We discuss in detail why in the case of a non-linear dependence between the configuration and a given property,
common restraining methods lead to unrealistic dynamics and based on that propose a modification of these methods.
The last chapter of this thesis gives an outlook on the development of analysis tools for MD data. We discuss a personal choice of questions and challenges the field will face in the coming years and also try to envision how sophisticated analysis tools could help to improve our understanding of biomolecular processes
Machine learning based prediction of esterases' promiscuity
No full textEls enzims són de gran interès per a la majoria de les indústries, no obstant la seva caracterització en el laboratori és costosa i molt laboriosa, fet que ha impulsat el desenvolupament de tecnologies de predicció de les activitats dels enzims. Malgrat això, els enzims industrials han de tenir unes propietats molt específiques com per exemple alta especificitat, alta activitat en condicions no biològiques i alta promiscuitat, característiques que no estan ben cobertes per les eines de predicció actuals. Per aquest motiu, amb aquest projecte, s'intenta mitigar el problema creant classificadors binaris que poden predir la promiscuitat de les esterases.Enzymes are of great interest for a vast variety of industries; however, the experimental characterization is very time consuming and expensive. Moreover, industrial enzymes need to adapt to nonbiological conditions while maintaining high activity, promiscuity and stereo-selectivity, properties that are not well covered, currently, by prediction technologies which means that their characterization still relies solely on experimentation. This project has the intention of mitigating the problem by developing binary classifiers and multi-classifiers that can predict the promiscuity of esterases, one of the many industrially relevant enzymes
- …
