I. Introduction
Tumorigenesis is the result of a complex interplay between multiple biological pathways. Therefore, use of only one single omics data to explore tumor progression will miss complex models that involve variation across multiple levels of biological regulation [1]. The development of high-throughput genomic technologies has made it possible for researchers to gain insights within multiple dimensions of genomic data. For example, The Cancer Genome Atlas (TCGA) project generates multidimensional genomic data, including gene expression (GE), DNA methylation (DM) and microRNA expression (ME) for the same cohort of tumor samples [2]. Unfortunately, advances in omics data analysis are far behind the fast accumulation of data. Identifying multidimensional regulatory modules (md-modules) from multidimensional genomic data is vital for investigation of the complex regulatory mechanisms by which elements at different levels interact with each other in biological systems.