Integration of high throughput data to detect groups of glioblastoma multiforme

The discovery of new subtypes of glioblastoma (Gb), grade IV glioma, based on high throughput molecular data, such as gene-expression and methylation microarrays, have become a challenging research eld for the scienti c community dedicated to neuro-oncology. The current diagnostic tools are not accurate enough and molecular signatures from high throughput data holds great promise to improve the patient's outcome. The detection of such sigatures requires a detailed statistical analysis of data to ensure their robustness. This study has compared the ablity to detect reliable signatures by separately using three different types of high throughput data (gene-expression, methylation and microRNA) and by integrating all the data types mentioned. A resampling approach was carried out based on multiple factor analysis and consensus clustering.. The objective of this study is to find signatures using different types of omics data in order to identify the most likely subtypes of glioblastomas (Gb). This study will be performed using data downloaded from the TCGA database, mainly consisting on 3 datasets: Gene expression, Methylation and miRNA microarrays. The application of appropriate methos will allow to achieve an improvement of the results respect to the results obtained using only gene expression microarray data