Statistical tools for classification, interpretation and prediction of biological data

Identifier:  TDX:2670

Handle: https://hdl.handle.net/20.500.11797/TDX2670

Authors:  Senan Campos, Oriol

Abstract:
Omics technologies arise a new systemic approach towards biology. We already know the genome of many species, and an approximated number of proteins, but we do not know how many metabolites are present in an organism or in a biological sample. With the most suited technique for metabolite identification, usually only 20-30 metabolites are identified after hard manual work. To solve this problem, we have developed CliqueMS, that tackles one of the main bottlenecks for the annotation of metabolomic experiments, the correct grouping and annotation of the multiple signals produced by a metabolite. In another investigation of the thesis, we explore how to combine several sources of omics data. In a practical application we study the therapeutic effect of hibiscus, by analyzing the response in metabolism and in gene expression, after its ingestion. The last investigation included in this thesis tackles a complex biological process, thrombosis. We study how changes interpretation and prediction of platelet deposition by using different computational models. By this models we demonstrate that platelet deposition can be predicted by measuring platelet concentration, the vessel tissue and some other variables. This models can be used to predict variables that are very difficult to measure, as it is platelet deposition