Computational and statistical methods for integrated analysis of biomedical data

Resumen

During recent years, the new omics technologies have revolutionized the biomedical research paradigm, changing from studying few specific elements based on previous hypotheses to studying complete systems like the genome or the transcriptome, generating hypotheses from the data. This change has created the necessity of a new profile in the biomedical research, the bioinformatician or computational biologist, who combines knowledge about biology, informatics and statistics in order to analyse these huge amounts of data and to develop new analytical methods. In this context of massive data generation, different public repositories were created where researchers can submit the data generated in their studies with the aim of guaranteeing the reproducibility of their results and of doing the data usable in other retrospective studies. For the last years, the amount of stored data in public repositories has grown exponentially thanks to the lowering costs of the necessary technologies to generate them. One of the most used repositories is the Gene Expression Omnibus (GEO), maintained by the NCBI. GEO contain the data generated in all types of omics projects, including gene expression, methylation or DNA sequencing, among others. The availability of all these amounts of information offers an invaluable resource to generate and test hypotheses through the use and integration of these data. However, for that aim, proper statistical and computational methods for integrating information are necessary. Among the strategies to reanalyse public data is the meta-analysis, consisting on the combination of the results from different studies using proper statistical techniques with the aim of increasing the statistical power and resolving discrepancies between studies, among other applications. The main objective of this doctoral thesis has been the development of computational methods for the integration of heterogeneous data sets with the aim of analysing them in conjunction using meta-analysis and integrated analysis methods.