In biomedical research, large-scale profiling of gene expression has become routine and offers a valuable means to evaluate changes in onset and progression of diseases, in particular cancer. An overwhelming amount of cancer genomics data has become publicly available, and the complexity of these data makes it a challenge to perform data exploration, integration and analysis, in particular for scientists lacking a background in computational programming or informatics. Many web interface tools make these large datasets accessible but are limited to process large datasets. To accelerate the translation of genomic data into new insights, we provide a simple method to explore and select data from cancer genomic datasets to generate gene-expression profiles of subsets that are of specific genetic, biological or clinical interest.