|
Abstract:
|
Graphical models have been used in many scientific fields for exploration of conditional independence relationships for a large set of random variables. In pan-cancer network analysis, we have mixed type of data from multiple types of cancer, such as gene expression, mutation, and DNA methylation. In order to examine the similarity and difference among the genomic and cellular alterations across different types of cancer, we proposed a novel method to jointly estimate the network structure across multiple cancers and multiple types of data, such as those mixed with Gaussian, multinomial, and Poisson. The method also allows people to incorporate domain knowledge into network construction by restricting some links to be included in or excluded from the networks. As a result, we observed some interesting and sensible results in six types of cancer that have been either verified in the literature or worth to be further explored. To our knowledge, this is the first work for joint estimation of multiple mixed graphical models.
|
