Bioinformatics methods for the genomics and metabolomics analysis of immune-mediated inflammatory diseases
- Autores: Arnald Alonso Pastor
- Directores de la Tesis: Antonio Julia Cano (dir. tes.), Sara Marsal Barril (codir. tes.), Alexandre Perera Lluna (tut. tes.)
- Lectura: En la Universitat Politècnica de Catalunya (UPC) ( España ) en 2015
- Idioma: español
- Tribunal Calificador de la Tesis: Montserrat Vallverdú Ferrer (presid.), Oscar Yanes (secret.), Juan de Dios Cañete Crespillo (voc.)
- Materias:
- Enlaces
- Resumen
During the last decade, genomics have been widely used to the characterization of the molecular basis of common diseases. Genome-wide association studies (GWAS) have been highly successful in characterizing the genetic variation that influences human traits including the susceptibility to common diseases. In metabolomics, recent improvements of analytical technologies have enabled the analysis of complete metabolomic profiles. Using this approach, high-throughput metabolomics studies have already demonstrated a high potential for the discovery of disease biomarkers.
The use of powerful high-throughput measurement technologies has resulted in the generation of large datasets of biological variation. In order to extract relevant biological information from this data, highly specialized bioinformatics methods are required. This thesis is focused on the development of new methodological tools to improve the processing of genomics and metabolomics high-throughput data. These new tools have been used in the analysis framework of the Immune-Mediated Inflammatory Diseases (IMIDs) Consortium. The IMID Consortium is a large Spanish network of biomedical researchers on autoimmune diseases, which holds one of the largest collections of biological samples from this group of diseases, as well as healthy controls.
The first analysis tool that has been developed is a computationally efficient algorithm for simultaneous genotyping of single nucleotide polymorphisms (SNPs) and copy number variants (CNVs) using microarray data. This bioinformatics tool, called GStream, integrates the genotyping of both types of genomic variants into a single processing pipeline. We demonstrate that the developed algorithms provide a significant increase in genotyping accuracy and call rate when compared to previous algorithms. Using GStream, the researchers performing large-scale GWASs will not only benefit from the combined and fast genotyping of SNPs and CNVs but, more importantly, they will also improve the accuracy and therefore the statistical power of their studies.
The second tool that was developed during this thesis was FOCUS, a bioinformatics framework that provides a complete data analysis workflow for high-throughput metabolomics studies based on one-dimensional nuclear magnetic resonance (NMR). FOCUS workflow includes quality control, peak alignment, peak picking and metabolite identification. The algorithms included in FOCUS were designed to overcome several technical challenges that can dramatically affect the quality of the results. FOCUS allows users to easily obtain high-quality NMR feature matrices, which are ready for chemometric analysis, as well as metabolite identification scores for each peak that greatly simplify the biological interpretation of the results. When tested against previous NMR data processing methodologies, FOCUS clearly showed a superior performance, even in datasets with high levels of spectral unalignment.
The final research work included in this thesis is a GWAS in Crohn's disease (CD) clinical phenotypes. CD is the most prevalent chronic inflammatory disease of the bowel, and is characterized by segmental and transmural inflammation of the gastrointestinaltract. CD is a highly heterogeneous disease, with patients showing different degrees of severity. The identification of the genetic basis associated with disease severity is therefore a major objective in CD translational research. The present PhD thesis includes the first GWAS of clinically relevant phenotypes in CD. A total of 17 phenotypes associated with different clinical complications were analyzed. In this study, we identified new genetic regions significantly associated to complicated disease course, disease location, mild disease course, and erythema nodosum. These findings are of high relevance since they show the existence of a genetic component for disease heterogeneity that is independent of the genetic variation associated with susceptibility to CD.
