Despite some great success, many human diseases cannot be effectively treated, prevented or cured, yet. Moreover, prescribed drugs are often not very efficient and cause undesired side effects. Hence, there is a need to investigate the molecular basis of diseases and adverse drug reactions in more detail. For this purpose, relevant biomedical data needs to be gathered, integrated and analysed in a meaningful way. In this regard, we have developed novel integrative analysis approaches based on both perspectives, classical multivariate statistics and systems biology. A novel multilevel statistical method has been developed for exploiting molecular and pharmacological information for a set of drugs in order to investigate undesired side effects. Systems biology approaches have been used to study the genetic basis of human diseases at a global scale. For this purpose, we have developed an integrated gene-disease association database and tools for user-friendly access and analysis. We showed that modularity applies for mendelian, complex and environmental diseases and identified disease-related core biological processes. We have constructed a workflow to investigate adverse drug reactions using our gene-disease association database. A detailed study of currently available pathway data has been performed to evaluate its applicability to build network models. Finally, a strategy to integrate information about sequence variations with biological pathways has been implemented to study the effect of the sequence variations onto biological processes. In summary, the developed methods are of immense practical value for other biomedical researchers and can aid to improve the understanding of the molecular basis of diseases and adverse drug reactions.A pesar de que existen tratamientos eficaces para las enfermedades, no hay todavía una cura o un tratamiento efectivo para muchas de ellas. Asimismo los medicamentos pueden ser ineficaces o causar efectos secundarios indeseables. Por lo tanto, es necesario investigar en profundidad las bases moleculares de las enfermedades y de los efectos secundarios de los medicamentos. Para ello, es necesario identificar y analizar de forma integrada los datos biomédicos relevantes. En este sentido, hemos desarrollado nuevos métodos de análisis e integración de datos biomédicos que van desde el análisis estadístico multivariante a la biología de sistemas. En primer lugar, hemos desarrollado un nuevo método estadístico multinivel para la explotación de la información molecular y farmacológica de un conjunto de drogas a fin de investigar efectos secundarios no deseados. Luego, hemos usado métodos de biología de sistemas para estudiar las bases genéticas de enfermedades humanas a escala global. Para ello, hemos integrado en una base de datos asociaciones entre genes y enfermedades y hemos desarrollado herramientas para el fácil acceso y análisis de los datos. Mostramos que las enfermedades mendelianas, complejas y ambientales presentan modularidad e identificamos los procesos biológicos relacionados con dichas enfermedades. Hemos construido una herramienta para investigar las reacciones adversas a los medicamentos basada en nuestra base de datos de asociaciones entre genes y enfermedades. Realizamos un estudio detallado de los datos disponibles sobre los procesos biológicos para evaluar su aplicabilidad en la construcción de modelos dinámicos. Por último, desarrollamos una estrategia para integrar la información sobre las variaciones de secuencia de genes con los procesos biológicos para estudiar el efecto de dichas variaciones en los procesos biológicos. En resumen, los métodos presentados en esta tesis constituyen una herramienta valiosa para otros investigadores y pueden ayudar a mejorar la comprensión de las bases moleculares de las enfermedades y de las reacciones adversas a los medicamentos.

Whole-exome and genome sequencing are widely used to diagnose individual patients. However, despite its success, this approach leaves many patients undiagnosed. This could be due to the need to discover more disease genes and variants or because disease phenotypes are novel and arise from a combination of variants of multiple known genes related to the disease. Recent rapid increases in available genomic, biomedical, and phenotypic data enable computational analyses, reducing the search space for disease-causing genes or variants and facilitating the prediction of causal variants. Therefore, artificial intelligence, data mining, machine learning, and deep learning are essential tools that have been used to identify biological interactions, including protein-protein interactions, gene-disease predictions, and variant--disease associations. Predicting these biological associations is a critical step in diagnosing patients with rare or complex diseases. In recent years, computational methods have emerged to improve gene-disease prioritization by incorporating phenotype information. These methods evaluate a patient's phenotype against a database of gene-phenotype associations to identify the closest match. However, inadequate knowledge of phenotypes linked with specific genes in humans and model organisms limits the effectiveness of the prediction. Information about gene product functions and anatomical locations of gene expression is accessible for many genes and can be associated with phenotypes through ontologies and machine-learning models. Incorporating this information can enhance gene-disease prioritization methods and more accurately identify potential disease-causing genes. This dissertation aims to address key limitations in gene-disease prediction and variant prioritization by developing computational methods that systematically relate human phenotypes that arise as a consequence of the loss or change of gene function to gene functions and anatomical and cellular locations of activity. To achieve this objective, this work focuses on crucial problems in the causative variant prioritization pipeline and presents novel computational methods that significantly improve prediction performance by leveraging large background knowledge data and integrating multiple techniques. Therefore, this dissertation presents novel approaches that utilize graph-based machine-learning techniques to leverage biomedical ontologies and linked biological data as background knowledge graphs. The methods employ representation learning with knowledge graphs and introduce generic models that address computational problems in gene-disease associations and variant prioritization. I demonstrate that my approach is capable of compensating for incomplete information in public databases and efficiently integrating with other biomedical data for similar prediction tasks. Moreover, my methods outperform other relevant approaches that rely on manually crafted features and laborious pre-processing. I systematically evaluate our methods and illustrate their potential applications for data analytics in biomedicine. Finally, I demonstrate how our prediction tools can be used in the clinic to assist geneticists in decision-making. In summary, this dissertation contributes to the development of more effective methods for predicting disease-causing variants and advancing precision medicine.