A Conjugate Exponential Model for Cancer Tissue Heterogeneity

The diagnosis and treatment of cancer is made difficult by the heterogeneous nature of the cell population. Determining its compositional breakup from measurements of various measurable traits (such as gene expression measurements) is an important problem in the field of cancer diagnosis and treatment. In addition, the computational aspect of the problem also needs attention. The processing of the collected data must be done as efficiently as possible in terms of computational speed and memory requirements. The use of Markov chain Monte Carlo methods is time consuming, and hence, other methods need to be used for the analysis. In this paper, we develop a model for heterogeneous cancer tissue, which uses quantitative polymerase chain reaction gene expression data to determine the compositional breakup of cell populations in the heterogeneous tissue. We develop a fast algorithm for the model using variational methods and demonstrate its use on synthetic and real-world gene expression data collected from fibroblasts and compare the performance of the algorithm with other methods such as Markov chain Monte Carlo and expectation maximization.