Figure2: Local Analysis and optimization in six steps. 1: The first step is to construct a matrix of genomic localisation for the tested gene list. 2: Then the corresponding hazard distribution is computed.2: A global test is performed to detect potential regions of interest at a large scale. Then each base unit are tested if they are located in large regions of interest and if there are more genes than expected (real number > hazard estimation). 4: A step of optimization test if extension of statistically significant regions of interest brings better results. 5: The final test is done after optimization. 6: Closer significant regions of interest can be merged if the resulting larger region improves statistical significance.