Leibniz MMS Days 2018 - Abstract
Koenig, Rainer
Constructing an appropiate and specific gene regulatory network of a cellular patho-mechanism is key to identify the relevant regulatory interactions suiting for biomarkers and drug targets. Recently, we developed a constrained based model and embedded it into a machine learning procedure to predict gene regulators by selecting regulators which best predict expression of the gene of interest using large scale gene expression data and a comprehensive set of regulator binding information from databases. We now extended this by integrating a regulator – regulator interaction model optimized to gain a compact regulatory network employing modularity.Telomeres are at the ends of chromosomes and shorten with each cell division preventing unlimited proliferation. Cancer cells overcome this restriction by upregulating the telomerase. We apply our approach to study the regulation of the human telomerase in prostate cancer and identified regulators with, interesingly, high clinical relevance.