Using serious games for improving medical education: an application to cytopathology

El objetivo principal del proyecto es desarrollar una plataforma compuesta por aplicaciones educativas gamificadas para el entrenamiento de personal médico en países de recursos limitados en citopatología mediante dispositivos Android de bajo presupuesto. Antes de desplegar la plataforma en países con recursos limitados, va a ser probada en un curso de Introducción a Citopatología de la Escuela Médica de Harvard. El proyecto final debe funcionar tanto en PCs como en dispositivos Android de bajo coste (p.e. 50 dólares americanos, Amazon Kindle Fire 7 pulgadas) y no puede depender de una conexión a internet continua. Se han analizado algunas aplicaciones con propósito de juego y simulaciones gamificadas para tener una base de conocimiento común entre expertos médicos y desarrolladores. También se han estudiado juegos y aplicaciones cuyo objetivo es hacer uso de imágenes médicas para entrenamiento de personal médico o están enfocadas al diagnóstico mediante colaboración por parte de personal no-médico. Esto nos ha permitido identificar las mejores mecánicas de juego para nuestro caso de uso. A continuación, se han comparado diferentes herramientas de edición y motores de juegos desde el punto de vista del rendimiento ofrecido, las plataformas soportadas, su documentación y licencia. Todo ello nos ha permitido elegir la tecnología de desarrollo (libGDX). Finalmente, diseñamos e implementamos un sistema integrado de aplicaciones (editor de contenido y generador de juegos). El sistema está enfocado a reducir la dependencia entre el personal experto y los desarrolladores para crear y mantener contenido educativo. Se trata de una arquitectura formada por un servicio RESTful, y un editor asociado, orientado a la gestión de contenido educativo orientado para citopatología y dos clientes para diferentes plataformas (PC y Android) que consumen dicho servicio. Finalmente, se presentan las conclusiones y el trabajo futuro del proyecto.

Towards the Knowledge-based Design of Universal Influenza Epitope Ensemble Vaccines

Motivation: Influenza A viral heterogeneity remains a significant threat due to unpredictable antigenic drift in seasonal influenza and antigenic shifts caused by the emergence of novel subtypes. Annual review of multivalent influenza vaccines targets strains of influenza A and B likely to be predominant in future influenza seasons. This does not induce broad, cross protective immunity against emergent subtypes. Better strategies are needed to prevent future pandemics. Cross-protection can be achieved by activating CD8+ and CD4+ T cells against highly-conserved regions of the influenza genome. We combine available experimental data with informatics-based immunological predictions to help design vaccines potentially able to induce cross-protective T-cells against multiple influenza subtypes. Results: To exemplify our approach we designed two epitope ensemble vaccines comprising highlyconserved and experimentally-verified immunogenic influenza A epitopes as putative non-seasonal influenza vaccines; one specifically targets the US population and the other is a universal vaccine. The USA-specific vaccine comprised 6 CD8+ T cell epitopes (GILGFVFTL, FMYSDFHFI, GMDPRMCSL, SVKEKDMTK, FYIQMCTEL, DTVNRTHQY) and 3 CD4+ epitopes (KGILGFVFTLTVPSE, EYIMKGVYINTALLN, ILGFVFTLTVPSERG). The universal vaccine comprised 8 CD8+ epitopes: (FMYSDFHFI, GILGFVFTL, ILRGSVAHK, FYIQMCTEL, ILKGKFQTA, YYLEKANKI, VSDGGPNLY, YSHGTGTGY) and the same 3 CD4+ epitopes. Our USA-specific vaccine has a population protection coverage (portion of the population potentially responsive to one or more component epitopes of the vaccine, PPC) of over 96% and 95% coverage of observed influenza subtypes. The universal vaccine has a PPC value of over 97% and 88% coverage of observed subtypes.

Diagnóstico automático de patologías cardíacas de bloqueo

Los electrocardiogramas (ECG) registran la actividad eléctrica del corazón a través de doce señales principales denominadas derivaciones. Estas derivaciones son analizadas por expertos médicos observando aquellos segmentos de la señal eléctrica que determinan cada una de las patologías que pueden afectar al corazón. Este hecho en general, es un condicionante muy importante para el diseño de sistemas expertos de diagnóstico médico, ya que es preciso conocer, delimitar y extraer de la señal eléctrica aquellos segmentos que determinan la patología. Dar solución a estos problemas, sería fundamental para facilitar el diseño de sistemas expertos para el diagnóstico de enfermedades cardiacas. El objetivo de este trabajo es demostrar que es posible identificar patologías cardiacas analizando la señal completa de las diferentes derivaciones de los ECGs, y determinar puntos concretos que determinan la patología en lugar de segmentos de la señal. Para ello se ha utilizado una BBDD de electrocardiogramas y se ha determinado mediante un algoritmo, los puntos de la señal que determinan la patología. Se ha aplicado a la patología de bloqueos de rama y los puntos obtenidos con el algoritmo se han utilizado para el diseño de un clasificador automático basado en redes neuronales artificiales, obteniendo un coeficiente de sensibilidad del 100% y de especificad del 99.24%, demostrando su validez para el diseño de sistemas expertos de clasificación.

Epigenetics and alternative splicing

Dissertação de Mestrado, Oncobiologia: Mecanismos Moleculares do Cancro, Departamento de Ciências Biomédicas e Medicina, Universidade do Algarve, 2015

Automatic learning of pre-miRNAs from different species.

Discovery of microRNAs (miRNAs) relies on predictive models for characteristic features from miRNA precursors (pre-miRNAs). The short length of miRNA genes and the lack of pronounced sequence features complicate this task. To accommodate the peculiarities of plant and animal miRNAs systems, tools for both systems have evolved differently. However, these tools are biased towards the species for which they were primarily developed and, consequently, their predictive performance on data sets from other species of the same kingdom might be lower. While these biases are intrinsic to the species, their characterization can lead to computational approaches capable of diminishing their negative effect on the accuracy of pre-miRNAs predictive models. We investigate in this study how 45 predictive models induced for data sets from 45 species, distributed in eight subphyla/classes, perform when applied to a species different from the species used in its induction. Results: Our computational experiments show that the separability of pre-miRNAs and pseudo pre-miRNAs instances is species-dependent and no feature set performs well for all species, even within the same subphylum/class. Mitigating this species dependency, we show that an ensemble of classifiers reduced the classification errors for all 45 species. As the ensemble members were obtained using meaningful, and yet computationally viable feature sets, the ensembles also have a lower computational cost than individual classifiers that rely on energy stability parameters, which are of prohibitive computational cost in large scale applications. Conclusion: In this study, the combination of multiple pre-miRNAs feature sets and multiple learning biases enhanced the predictive accuracy of pre-miRNAs classifiers of 45 species. This is certainly a promising approach to be incorporated in miRNA discovery tools towards more accurate and less species-dependent tools.

Role of Ccbe1 during cardiac differentiation of mouse ESCs

Tese de Doutoramento, Ciências Biomédicas, Departamento de Ciências Biomédicas e Medicina, Universidade do Algarve, 2016

Descubrimiento de fármacos basado en cribado virtual refinado con enfoques neuronales paralelos

En el campo de la medicina clínica es crucial poder determinar la seguridad y la eficacia de los fármacos actuales y además acelerar el descubrimiento de nuevos compuestos activos. Para ello se llevan a cabo ensayos de laboratorio, que son métodos muy costosos y que requieren mucho tiempo. Sin embargo, la bioinformática puede facilitar enormemente la investigación clínica para los fines mencionados, ya que proporciona la predicción de la toxicidad de los fármacos y su actividad en enfermedades nuevas, así como la evolución de los compuestos activos descubiertos en ensayos clínicos. Esto se puede lograr gracias a la disponibilidad de herramientas de bioinformática y métodos de cribado virtual por ordenador (CV) que permitan probar todas las hipótesis necesarias antes de realizar los ensayos clínicos, tales como el docking estructural, mediante el programa BINDSURF. Sin embargo, la precisión de la mayoría de los métodos de CV se ve muy restringida a causa de las limitaciones presentes en las funciones de afinidad o scoring que describen las interacciones biomoleculares, e incluso hoy en día estas incertidumbres no se conocen completamente. En este trabajo abordamos este problema, proponiendo un nuevo enfoque en el que las redes neuronales se entrenan con información relativa a bases de datos de compuestos conocidos (proteínas diana y fármacos), y se aprovecha después el método para incrementar la precisión de las predicciones de afinidad del método de CV BINDSURF.

Genome-wide copy number variation (CNV) detection in Nelore cattle reveals highly frequent variants in genome regions harboring QTLs affecting production traits.

Background: Copy number variations (CNVs) have been shown to account for substantial portions of observed genomic variation and have been associated with qualitative and quantitative traits and the onset of disease in a number of species. Information from high-resolution studies to detect, characterize and estimate population-specific variant frequencies will facilitate the incorporation of CNVs in genomic studies to identify genes affecting traits of importance. Results: Genome-wide CNVs were detected in high-density single nucleotide polymorphism (SNP) genotyping data from 1,717 Nelore (Bos indicus) cattle, and in NGS data from eight key ancestral bulls. A total of 68,007 and 12,786 distinct CNVs were observed, respectively. Cross-comparisons of results obtained for the eight resequenced animals revealed that 92 % of the CNVs were observed in both datasets, while 62 % of all detected CNVs were observed to overlap with previously validated cattle copy number variant regions (CNVRs). Observed CNVs were used for obtaining breed-specific CNV frequencies and identification of CNVRs, which were subsequently used for gene annotation. A total of 688 of the detected CNVRs were observed to overlap with 286 non-redundant QTLs associated with important production traits in cattle. All of 34 CNVs previously reported to be associated with milk production traits in Holsteins were also observed in Nelore cattle. Comparisons of estimated frequencies of these CNVs in the two breeds revealed 14, 13, 6 and 14 regions in high (>20 %), low (<20 %) and divergent (NEL > HOL, NEL < HOL) frequencies, respectively. Conclusions: Obtained results significantly enriched the bovine CNV map and enabled the identification of variants that are potentially associated with traits under selection in Nelore cattle, particularly in genome regions harboring QTLs affecting production traits.

Avaliação do desempenho relativo de bancos de dados NoSQL para arquivos de genótipos.

A bioinformática e a genômica trabalham com bases de dados fora do padrão tradicional ou clássico que, por sua vez, caracterizam-se pela organizacão tabular e pelo tratamento destas em SGBDRs. Arquivos de genótipos são exemplos de bases de dados não clássicas e são caracterizados por serem gerados como arquivos textos, com dados desbalanceados, com alta dimensionalidade e por ocuparem muito espaço, entre outros aspectos. Os SGBDRs não têm se mostrado uma boa solucão para o tratamento de tais bases e, portanto, o presente trabalho busca avaliar o desempenho relativo entre bancos de dados NoSQL que representam duas famílias de diferentes modelo de dados, a partir de cenários de teste para a manipulação de arquivos de genótipo.

Regulação de proteínas da via glicolítica em raízes de Elaeis guineensis Jacq. acometidas pelo amarelecimento datal.

O aparecimento do Amarelecimento Fatal (AF) tem sido um empecilho para o crescimento da cultura do dendezeiro. Neste sentido, várias pesquisas foram realizadas visando identificar as causas do AF. No entanto outras estratégias precisam ser aplicadas. Técnicas em proteômica vêem se modernizando para obter o perfil protéico qualitativo e quantitativo com maior velocidade e precisão. Sabendo que a regulação de componentes do metabolismo primário está envolvida direta ou indiretamente na defesa, o objetivo deste trabalho foi utilizar a cromatografia líquida bidimensional e espectrometria de massas (2D-UPLC/MSE) com o auxílio de ferramentas de bioinformática na identificação de alterações de proteínas envolvidas no metabolismo energético em raízes de plantas com AF. Os resultados revelaram a expressão diferencial de proteínas envolvidas na glicólise, onde a maior frequência de regulação positiva dessas em plantas assintomáticas pode estar relacionada ao retardo no desenvolvimento desta doença.

Detection of human interchromosomal trans-splicing in sequence databanks.

Trans-splicing is a common phenomenon in nematodes and kinetoplastids, and it has also been reported in other organisms, including humans. Up to now, all in silico strategies to find evidence of trans-splicing in humans have required that the candidate sequences follow the consensus splicing site rules (spliceosome-mediated mechanism). However, this criterion is not supported by the best human experimental evidence, which, except in a single case, do not follow canonical splicing sites. Moreover, recent findings describe a novel alternative tRNA mediated trans-splicing mechanism, which prescinds the spliceosome machinery. In order to answer the question, ?Are there hybrid mRNAs in sequence databanks, whose characteristics resemble those of the best human experimental evidence??, we have developed a methodology that successfully identified 16 hybrid mRNAs which might be instances of interchromosomal trans-splicing. Each hybrid mRNA is formed by a trans-spliced region (TSR), which was successfully mapped either onto known genes or onto a human endogenous retrovirus (HERV-K) transcript which supports their transcription. The existence of these hybrid mRNAs indicates that trans-splicing may be more widespread than believed. Furthermore, non-canonical splice site patterns suggest that infrequent splicing sites may occur under special conditions, or that an alternative trans-splicing mechanism is involved. Finally, our candidates are supposedly from normal tissue, and a recent study has reported that trans-splicing may occur not only in malignant tissues, but in normal tissues as well. Our methodology can be applied to 5'-UTR, coding sequences and 3'-UTR in order to find new candidates for a posteriori experimental confirmation.