Encode for cancer genomics to predict gene expression

I am just beginning start my first ever project by using the extended gene definition provided in the dataset of Encode for cancer genomics to predict gene expressions. I would be incredibly grateful if there could be an explanation about the layout of the text files. I have been unsuccessfully trying to understand how the extended gene was used to interpret the mutations and expression changes in the published article.

Thanks for your interest in the research and the extended gene annotation. We are preparing BED-formatted extended gene annotation and they will be available soon on our project website (http://encodec.encodeproject.org/). We will keep you informed.