miXGENE.ORG - a public tool for integrated analysis of microarray, microRNA and methylation data

Systems biology focuses on complex interaction within biological systems. These complex interactions, hidden both in high-throughput measurements and a vast amount of existing biological knowledge, are difficult to be mined without an aid of automated tools. MiXGENE is a workflow management tool that develops and analyzes accurate, mainly disease predictive, models from the raw omics data.

This project develops a web tool that facilitates integrated knowledge discovery from raw mRNA, miRNA and methylation data with concurrent utilization of the structured genomic background knowledge. There are two main project outputs: the workflow management tool (and the methodology and dedicated algorithms behind it) itself and the particular results reached in cooperation with clinical and biological departments working in the fields of myelodysplastic syndrome and germ cell tumors. The solution is based on relational learning algorithms, stochastic optimization, statistics and development of web applications. The tool outputs namely 1) biologically understandable patterns having a form of sets or annotated networks of specific related elements such as genes, proteins, miRNA sequences or methylation islands interconnected with particular subsets of biological samples under study and 2) predictive models classifying samples characterized by measurable molecular markers with unknown phenotypes.

Researchers involved: Jiří Kléma, Michal Anděl, Pavel Strnad

Official project website: http://mixgene.felk.cvut.cz/